Charged Droplet Dynamics in the Submicrometer Size Range - The

970

J. Phys. Chem. B 2009, 113, 970–976

Charged Droplet Dynamics in the Submicrometer Size Range Christopher J. Hogan, Jr.,*,†,‡ Pratim Biswas,† and Da-ren Chen† Department of Energy, EnVironmental, and Chemical Engineering, Washington UniVersity in St. Louis, Missouri, and Department of Mechanical Engineering, Yale UniVersity, New HaVen, Connecticut ReceiVed: September 1, 2008; ReVised Manuscript ReceiVed: NoVember 11, 2008

The highly charged droplets produced by electrospray are of significant importance in analytical chemistry, particularly in mass spectrometry. In spite of this, little is known about the mechanisms of charge loss for highly charged droplets in the submicrometer size range. In this study, electropsrayed droplet charge loss dynamics in the submicrometer size range were examined for the first time using tandem differential mobility analysis (TDMA) of aerosol particles originally enclosed in methanol-water droplets. After complete droplet evaporation, the remaining aerosol particles carried the nonejected charges from droplets; thus, particle electrical mobility spectra are reflective of the electrical mobility spectra of droplets of similar size. Bayesian data inversion was used to compare measured electrical mobility spectra to models for droplet losing charge by Coulombic fission and by single ion emission. Data inversion showed that, at a size of ∼40 nm, a transition of the mechanism of charge loss occurred. Methanol-water droplets larger than this size lost charge through Coulombic fissions at the Rayleigh limit with an effective surface tension of 0.050 N m-1, while droplets in the 10-40 nm size range lost charge through ion emission, maintaining a constant electric field of 1.1 V nm-1 on their surface during ion emission. This transition was observed experimentally for the first time in this study and is in good agreement with theoretical predictions from prior study of ion emitting nanodroplets. The data presented here provide the necessary link between studies of highly charged micrometer sized droplet and those of highly charged liquid clusters. Introduction Highly charged droplets, produced by electrospray, have been a source of scientific fascination for well over a century.1,2 In the past several decades, use of such droplets as means to aerosolize and ionize low volatility and high molecular weight compounds for analysis with mass spectrometers has become widespread.3 The dynamics of highly charged droplets, i.e. the mechanism by and rates at which droplets lose charge as they evaporate, are critical to analyte ionization in electrospray ionization (ESI), strongly influencing the charge distribution on the ionized analytes4,5 as well as the ion production rate.6,7 Due to the analytical importance of charged droplets, study of their dynamics, from droplet formation via breakup of charged liquid jets8,9 to the final stages of the evaporation process,10 is of interest. The original size and charge of droplets formed by electrospray has been examined experimentally11-13 and theoretically.14-16 As these droplets evaporate, the repulsive electrostatic forces on their surfaces increase to a point at which the droplets are no longer stable. Droplets which are unstable due to electrostatic repulsion undergo Coulombic fission, in which they eject mass and charge in the form of smaller progeny droplets. Further evaporation of both the original parent droplet and progeny droplets drives subsequent Coulombic fissions. For highly charged droplets in the micrometer size range, single droplet light scattering techniques may be used for their measurement and the fission process can be modeled as a quasistatic process; thus the fission process has been studied experimentally17-20 and theoretically21,22 in this size range. At * To whom correspondence should be addressed. Tel.: 203-432-4353. Fax: 203-432-7654. E-mail: [email protected]. † Washington University in St. Louis. ‡ Yale University.

the other end of the spectrum, when a multiply charged droplet is several nanometers in diameter, the electric field on the droplet surface is large enough to drive the emission of charge carrying molecules and molecular clusters from the droplet. This ion evaporation or emission23,24 process differs from a Coulombic fission in that during ion emission the original droplet does not eject large fractions of its total charge in discrete events, but rather single ions in a semicontinuous manner. Molecular scale simulations25-27 have been used to investigate ion emission and droplet dynamics in the nanometer size range. While direct observation of nanometer sized droplets is hindered by the rapid time scales on which they evaporate, nanoparticles originally enclosed within charged nanodroplets will retain the droplet charge at the moment solvent evaporation completes,28 allowing for indirect observation. Using mobility10,29-31 and mass spectrometry4,5 of multiply charged clusters and macromolecules, nanosized charged droplets have also been studied experimentally. Micrometer sized droplets will inevitability evaporate to submicrometer sizes (here we consider the submicrometer size range to be from ∼50 nm up to 1 µm in diameter). Likewise, nanometer sized droplets typically derive from submicrometer droplets. In modeling droplet fissions, is it commonly assumed that the characteristic electrical relaxation length is smaller than the droplet diameter.22 This assumption breaks down in the submicrometer range, and submicrometer systems are typically too large for examination by molecular scale simulations. Experimentally, highly charged submicrometer droplets have also been somewhat inaccessible; most mass spectrometers are incapable of detecting objects in this mass range,32,33 and intensity of scattering light by submicrometer droplets is low. Clearly, the lack of information on submicrometer droplet dynamics prevents complete description of the electrospray process. In ESI, the nature of the Coulombic fission process in

10.1021/jp807765n CCC: $40.75  2009 American Chemical Society Published on Web 01/05/2009

Submicrometer Charged Droplet Dynamics

J. Phys. Chem. B, Vol. 113, No. 4, 2009 971

Figure 1. Schematic representation of the tandem differential mobility analysis (TDMA) system used for measurement of electrosprayed particles.

the submicrometer size range strongly influences the efficiency of analyte ionization.7,34 The charge distribution on extremely large megaDalton and gigaDalton analytes in ESI35,36 are determined by submicrometer droplet dynamics, as the analytes themselves approach submicrometer dimensions. Charged particle deposition techniques37-39 and single particle image detection40 make use of submicrometer particles which are charged via electrospray. In addition, a transition in the mechanism of charge ejection, from Coulombic fission to ion emission, must occur in this size range.5 Here, we report on experimental measurements of highly charged methanol-water droplet dynamics in the submicrometer size range using tandem differential mobility analysis (TDMA).41,42 As has been used in study of nanometer sized droplets, the electrical mobility distributions of submicrometer particles and residues were used to determine the charge limits on submicrometer droplets. Due to the modest resolution of the differential mobility analyzers (DMAs) utilized, a data inversion routine43 was used to compare measured data to simple models of Coulombic fission and ion emission. A transition from charge loss by Coulombic fission to charge loss by ion emission was clearly identifiable. To the best of our knowledge, this is the first examination of electrosprayed droplets at the submicrometer scale and links studies of micrometer sized droplets to those of small nanometer droplets. Experimental Methods Electrospay Procedure. Submicrometer droplet dynamics were studied by electrospraying methanol-water suspensions and solutions of nanoparticles and nonvolatile species, respectively, and allowing droplets to dry completely, leaving residual aerosol particles. The tested particles were 500 nm SiO2 (Silica, Duke Scientific, Palo Alto, CA), 426, 200, 100, and 80 nm polystyrene latex (PSL, Duke Scientific), ∼90 nm silver (SigmaAldrich, Saint Louis, MO), and 50 nm gold (Ted Pella Inc., Redding, CA). Sucrose and ∼8000 Da polyethylene glycol (PEG) were also used to make electrospray solutions which would give rise to multiply charged aerosol particles upon drying. Electrosprayed suspensions and solutions were either 50:50 methanol:water or 99:1 methanol:water by volume. Pure water was not used due to formation of a corona discharge when

electrospraying pure water, which lead to nonrepeatable results. All tested particles either were already dispersed in an aqueous suspension, or rapidly agglomerated in pure methanol;44 thus, pure methanol was also not used for electrospraying. Either ammonium acetate (NH4Ac, Sigma Aldrich) or triethylammonium bicarbonate (TEAB, Sigma-Aldrich) was added to suspensions and solutions to increase the electrical conductivity. Sucrose, PEG, and methanol were all purchased from SigmaAldrich, while water was deionized and filtered using a Milli-Q Advantage System (Millipore, USA). A schematic representation of the experimental system is shown in Figure 1. A syringe pump (Harvard Apparatus) was used to control the flowrate of electrospray solutions and suspensions in the 1.00-4.00 µL min-1 range. The electrospray capillary was a tapered stainless steel tube, with an inner diameter of 85 µm and an outer diameter of 110 µm. A high voltage DC power supply (Spellman HV, Hauppauge, NY) was used to create a potential difference of ∼2-3 kV between the electrospray capillary and the surrounding stainless steal chamber. The electric current of charged particles and ions depositing onto the chamber walls was monitored using an ammeter (Keithley Instruments Inc., Cleveland, OH); stable current readings in the 100-300 nA range implied that the electrospray was operating in the cone-jet mode.12 Particle-free air at a flowrate of 5.0 L min-1 was used to carry electrospray droplets through a 0.3125 cm diameter opening at the bottom of the electrospray chamber and into the inlet region of a nanoDMA (model 3085, TSI Inc., Minneapolis, MN). On the basis of the electrical conductivities and electrospray flowrates used in this study,45 electrosprayed droplets would have been in the micrometer size range, which is large enough to ensure that, prior to complete evaporation, the droplets either underwent fission at least once or emitted ions, and small enough to ensure that droplets evaporated completely prior to entering the nanoDMA classification region.46 Unlike previous studies of charged droplet dynamics,17-20 the region where droplet evaporation occurs was relatively free of electric potential differences in this study. Tandem Differential Mobility Analysis. After complete droplet evaporation, particles, both colloidal and from the drying of nonvolatile molecules, were aerosolized and retained the

972 J. Phys. Chem. B, Vol. 113, No. 4, 2009

Hogan et al.

charge that was on the droplet just prior to complete droplet evaporation.4,5,29-31 The use of TDMA to measure the electrical mobility distribution of size classified aerosol particles is described briefly here and can be found in greater detail elsewhere.43,47 Readers not concerned with the technical details of TDMA may proceed to the subsequent Data Inversion section, without much loss in scope. Dried, highly charged aerosol particles with unknown size and charge were drawn into the nano-DMA at a flowrate of 1.5 L min-1, and the nano-DMA was operated with a recirculating sheath flow of air at 18.0 L min-1. DMAs act as electrical mobility filters,48 only allowing transmission of particles with a specified electrical mobility, Zp, given by the following equation:

Zp(n, dp) )

neCc(dp) 3πµdp

(1)

where n is the integer number of charges on the particle, e is the unit electron charge (1.6 × 10-19 C), dp is the particle diameter, µ is the gas dynamic viscosity, and Cc is the semiempirical Cunningham correction factor, given as

Cc ) 1 +

(

( ))

dp 2λ a + b exp -g dp λ

be determined precisely from the electrical mobility distributions, using the known values of Zp1 and dp2 with eq 1. However, the DMAs used here were of modest resolution, and transmitted particles with a narrow range of electrical mobilities. For highly charged particles, the change in electrical mobility associated with a change in the number of charges on a particle from n to n ( 1 was not detectable by the DMA. To further complicate matters, the size of particles transmitted through the second DMA was also dependent on the resolution of the DMA. To account for DMA resolution and to compare measured mobility spectra to expected results based on models of ESI, a Bayesian inversion routine43 was employed. The inversion routine requires input of the DMA transfer function,51 as well as a proposed charging model, in this case a model for highly charged droplet dynamics. The nondiffusing DMA transfer function was used here, which is correct for most of the size range studied. For the smaller particles examined (10 nm), small errors may have been introduced by not accounting for diffusion,52 but these errors are much smaller than the inherent variation in measured data. Data were compared to two models for the dynamics of highly charged droplets. The first model was for Coulombic fissions. Droplets undergo Coulombic fission at the Rayleigh limit, described by the following equation:

(1a)

where λ is the mean free path of gas molecules (66.5 nm for N2 at standard temperature and pressure), and a, b, and g are constants with values of 1.165, 0.483, and 0.4985, respectively.49 Particles with known electrical mobility, Zp1, but still unknown n and dp were transmitted through the nano-DMA. At the nanoDMA outlet, bipolar ions, produced by a Kr-85 R-source (TSI Inc.), were used to reduce aerosol particle charge such that the charge distribution on transmitted aerosol particles followed the steady-state Fuchs distribution.50 In the size range examined, 10-500 nm, the number of charges on aerosol particles after bipolar charging was significantly lower than would be expected after electrospray ionization, with a large fraction of particles being uncharged. Of the remaining charged particles, most particles, particularly the smaller particles examined, were singly charged (n ) 1).12 After bipolar charging, particles were sent through a second DMA, either a long-DMA (TSI model 3081) or a nano-DMA, operated with a recirculating sheath flowrate between 5.0 and 10.0 L min-1, which was set to transmit particles of electrical mobility Zp2. In all cases, Zp2 , Zp1, with Zp2 being intentionally set to the electrical mobility of a particle with n ) 1, and dp2 in the 10-500 nm size range (Zp2 and dp2 are related through eq 1 with n ) 1). Therefore, the second DMA acted as a particle size filter, and particles transmitted through the TDMA system originally had electrical mobility Zp1 and diameter dp2. Aerosol particles at the outlet of the second DMA were detected using a condensation particle counter (CPC, model 3025, TSI Inc.), in which butanol vapor was condensed onto particles, growing them to micrometer sizes and allowing them to be detected by light scattering. To collect electrical mobility spectra at a specific particle size, Zp2 was fixed as Zp1 was varied. For each tested Zp1 and Zp2 combination, the CPC was operated in continuous counting mode for 3 min. At least eight values of Zp1 were used to define each electrical mobility spectrum. All measurements were repeated in triplicate to ensure repeatability. Data Inversion. With ideal electrical mobility filters of high resolution, the charge distribution on measured particles could

nR )

π (8ε σD 3)1/2 e 0 D

(2)

where nR is the critical number of charges on the droplet, ε0 is the permittivity of free space (8.854 × 10-12 F m-1), σ is the effective surface tension of the droplet, and DD is the droplet diameter. When eq 2 is satisfied (by a decrease in DD through evaporation), the droplet will lose a fraction of its charge, φ, during the fission event. If the final size at the end of solvent evaporation is the particle diameter dp, droplets lose charge only through fission events, droplets undergo at least one fission prior to completely evaporating, and the original droplet size distribution is not completely monodisperse (this is true in electrospray, where droplet size distribution functions have geometric standard deviations typically between 1.05-1.20 in the cone-jet mode13); then, it follows that the charge on a electrosprayed particle of diameter dp will be approximately uniformly distributed, with an upper limit of nR(dp) and a lower limit of (1 - φ)nR(dp). The unknown parameters in this charging model are then σ, the effective droplet surface tension near the end of evaporation, and φ, the fraction of charge lost during a fission event. Using the Bayesian inversion routine, test electrical mobility spectra were constructed for trial values of σ and φ and quantitatively compared to measured spectra. Mean values and standard deviations for σ and φ were determined by comparing multiple test spectra to measured spectra. The second model for charge loss of highly charged droplets was for ion emission. The kinetics of ion emission show that the electric field, E* on the surface of an ion emitting droplet should be roughly independent of the droplet size,4,5,10 thus the following relation can be written:

nIE )

πε0E*DD2 e

(3)

where nIE is the number of charges on an ion emitting droplet. For the ion emission model, the charge on particles at the end of the evaporation process should be nIE.4 To compare the ion


Figure 2. Expected electrical mobilities for droplets at the Rayleigh limit with σ ) 0.073 and 0.022 N m-1 and droplets with constant surface electric field strengths of 1.0 and 2.0 V nm-1.


Figure 3. Representative electrical mobility spectra from TDMA.

emission model to experimental results, test spectra with trial values of E* were constructed and compared to measured spectra. As with unknown parameters in the Coulombic fission model, a mean value and standard deviation for E* were calculated by comparing multiple test spectra to measured spectra. Results and Discussion Expected and Raw Data. Due to the necessity of data inversion, it is useful and pedagogical to examine expected results for electrical mobility measurements. Figure 2 shows the electrical mobility as a function of size at the Rayleigh limit of pure methanol and pure water droplets (σ ) 0.022 and 0.073 N m-1), as well as the electrical mobility for ion emitting droplets with E* ) 1.0 and 2.0 V nm-1. For larger sizes, the electrical mobilities of fissioning droplets are less than those of ion emitting droplets; hence, Coulombic fissions should control droplet charging at large sizes. Conversely, the electrical mobilities of ion emitting droplets are less than those of fissioning droplets at smaller sizes. As a droplet evaporates, a transition from fission governed charging to ion emission is expected to occur,5 though the onset of ion emission has not yet been observed experimentally. It is interesting to note that at small sizes, the electrical mobility of a droplet scales with the square of the droplet diameter;30 while at slightly larger sizes electrical mobility scales with ∼DD1.5. Therefore, with certain combinations of E* and σ, measured electrical mobilities may be approximately size independent. Selected electrical mobility spectra for 30 nm sucrose particles, 50 nm gold particles, and 200 nm PSL particles are shown in Figure 3. The electrical mobility of max intensity in each spectrum decreased with increasing particle size. Without data inversion, it is already apparent that measurements do not agree with the ion emission model, in which the electrical mobility should increase with increasing size in the submicrometer range. However, conclusions beyond this qualitative observation, e.g. the effective surface tension of methanol-water mixtures at the end of evaporation and the size corresponding to the onset of ion emission, require data inversion. Electric Field and Surface Tension. TDMA and data inversion allow for determination of σ, φ, and E* as a function of size, i.e. the size filtered by the second DMA. Table 1 summarizes all measured data in this study, including the electrospray settings (solvent composition, flowrate, and electri-

Figure 4. Surface electric field strengths (a) and effective surface tensions (b) from data inversion for 50% methanol, 50% water, 1 mM ammonium acetate electrospray droplets.

cal conductivity) for each spectrum, particle size that the second DMA was set to transmit, particle chemical composition, the effective surface tension and percent charge loss from data inversion (( one standard deviation) with the Coulombic fission model, and the electric field strength at the particle surface (( one standard deviation) from data inversion with the ion emission model. Parameters from the inversion routine are presumably representative of droplet properties moments before evaporation is complete, i.e. approximately at the measured size. Should charge loss in droplets occur through Coulombic fissions, then the effective surface tension for a given set of electrospray conditions should be independent of size. Likewise, should


Figure 5. Electrospray governing curve for 50% methanol, 50% water, 1 mM ammonium acetate constructed from TDMA data.

charge loss occur via ion emission, then the surface electric field from data inversion should be size invariant. Figures 4a and 4b show the surface electric field and the square root of the effective surface tension as a function of size, taken from data inversion with electrical mobility spectra from 50% methanol, 50% water, 1 mM NH4Ac electrospray solutions and suspensions (described in the first 14 rows of Table 1). There was considerable variation in both the surface electric field and effective surface tension, although this variation did not appear to be influenced by the chemical composition of the particles used in measurement. The lack of material dependence on results was expected, based on the assumptions that droplet discharge dynamics determine the final charge states of nanoparticles and that particles were relatively spherical (their size corresponds to the droplet size at the end of evaporation). This is certainly true for electrosprayed nanoparticles, but polymer molecules may adopt a variety of conformations in the gas phase after evaporation completes. Previously, it was found that single chains of PEG aerosolized and charged via electrospray may adopt highly stretched, nonspherical configurations.53 However, in this study, residue PEG particles were composed of many polymer molecules. Electron microscopy has shown that multimolecule polymer particles, when produced via electrospray, are typically spherical.13,54 In spite of high data variability, the surface electric field appeared constant at sizes below ∼40 nm at a value of 1.1 V nm-1, similar to electric field strength found by Suh and coworkers31 using electrosprayed gold nanoparticles with methanol-water mixtures. Below ∼40 nm, the effective surface tension appeared to increase with increasing particle size, which is also expected for ion emitting droplets. Above 40 nm, the surface electric field steadily decreased and the surface tension showed no apparent variation with size, with a mean value of 0.050 N m-1 in the 40-500 nm range. In total, these data imply that droplets from electrosprayed 50% methanol, 50% water mixtures lose charge by Coulombic fission with an effective surface tension of 0.050 N m-1 above ∼40 nm, while below ∼40 nm, droplets lose charge by ion emission with E* ) 1.1 V nm-1 (note that the surface tension values from data inversion at sizes below 40 nm are therefore not the actual surface tensions of these droplets, as they do not appear to lose charge by Coulombic fission). ESI Governing Curve. Using E* and nIE to calculate charge as a function of size below 40 nm, and σ, φ, and nR for calculation above 40 nm, a complete curve for ESI can be contructed and is shown in Figure 5. Data points at and below 40 nm represent nIE with error bars corresponding to the standard

Hogan et al. deviation in E*. Above 40 nm, data points represent the quantity (1 - φ/2)nR and errors bar show the span from nR to (1 - φ)nR (the maximum and minimum charge on Rayleigh limited droplets). The data from Suh et al.31 are also shown, as are reference lines corresponding to E* ) 1.1 V nm-1 and σ ) 0.050 N m-1. Plotted in this manner, the dynamics of submicrometer droplets are overwhelmingly clear, and the transition from charge loss by Coulombic fission to charge loss by ion emission occurs at ∼40 nm for methanol-water mixtures. To the best of our knowledge, this is the first experimental evidence of this transition, which has been predicted in the original24 as well as more recent studies on ion emission.5 As methanol is more volatile than water, near the end of the evaporation process, droplets should have consisted primarily of water molecules, and the charge on particles would be expected to be similar to that on particles electrosprayed from purely aqueous solutions and suspensions.55 However, the effective surface tension of 0.050 N m-1 is much lower than that of pure water (0.073 N m-1), and the determined value of E* is lower than previously determined E* values for charge carriers in water.4,5 Data imply that some methanol must remain in the droplet throughout the evaporation process and that methanol is on the surface of the droplet, reducing the surface tension and associating with charge carrying species (reducing E*). In air saturated with methanol vapor, by Henry’s law droplets may be up to 6% methanol in volume. Although methanol evaporation from ESI droplets is not sufficient to saturate the surrounding air with methanol, this does indeed show that droplets could have sufficient methanol content for methanol to influence their properties. Both experimental56 and numerical studies57,58 have shown that aqueous-organic nandroplets often have a core-shell structure, with an inner core of water surrounded by an outer organic layer. Evaporating, heterogeneous, submicrometer droplets may exhibit a similar structure and, thus, may have a surface tension lower than that of bulk water. Charge Loss and Spray Condition Influences. For the same set of data as Figure 5, Figure 6 shows percent of droplet charge lost during explosion as a function of size (error bars represent the standard deviation from data inversion). A dashed line indicates the position of the transition point between Coulombic fission and ion emission, i.e. data points at sizes less than 40 nm are from ion emitting droplets, thus the percent charge loss from data inversion is not reflective of the amount of charge lost during an actual droplet fission. The percent charge loss decreased with increasing size in the 50-500 nm range. The decrease was slight, such that the absolute number of charges lost during droplet fission still increased with increasing size. Theoretical models21,22 of droplet fissions, which provide scaling relations for the current carried during fission and the size of the produced progeny droplets, are dependent on the assumption that the fission is a steady process; hence, the characteristic time for fission to occur must be much greater than the electrical relaxation time of the electrospray solution. For the solvents used here, and for most solvents amenable to electrospray, the characteristic time for fission and the electrical relaxation time are of the same order of magnitude for submicrometer droplets. Clearly, future modeling and experimental efforts are necessary for submicrometer scale charge loss to be better understood. The remaining measured electrical mobility spectra correspond to examination of the influences of initial droplet size (based on the electrical conductivity and electrospray flow rate12), the electrical conductivity, type of solute used, and volume fraction of water and methanol used in electrospray



TABLE 1: Summary of Experimental Conditions and Results from Data Inversion for All Measured Electrical Mobility Spectraa calculated results

solvent 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 1 mM NH4Ac 50% methanol, 50% water, 0.1 mM NH4Ac 50% methanol, 50% water, 4 mM NH4Ac 50% methanol, 50% water, 10 mM NH4Ac 50% methanol, 50% water, 4 mM TEAB 50% methanol, 50% water, 4 mM TEAB 50% methanol, 50% water, 4 mM TEAB 50% methanol, 50% water, 4 mM TEAB 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 1 mM NH4Ac 99% methanol, 1% water, 0.1 mM NH4Ac 99% methanol, 1% water, 4 mM NH4Ac 99% methanol, 1% water, 10 mM NH4Ac a

particle size [nm]

syringe flowrate [µL min-1]

electrical conductivity [µS cm-1]

surface tension [N m-1]

% charge loss

electric field [V nm-1]

sucrose

10

2.00

69.13

0.019 ( 0.005

48.4 ( 2.9

1.20 ( 0.28

sucrose

20

2.00

69.13

0.039 ( 0.009

48.2 ( 3.9

1.07 ( 0.28

sucrose

30

2.00

69.13

0.054 ( 0.014

47.1 ( 4.0

1.05 ( 0.28

sucrose

40

2.00

69.13

0.072 ( 0.019

48.8 ( 4.4

1.08 ( 0.25

gold

50

2.00

69.13

0.057 ( 0.015

51.6 ( 4.7

0.86 ( 0.23

polyethylene glycol (PEG)

65

2.00

69.13

0.026 ( 0.006

49.3 ( 3.0

0.49 ( 0.12

polystyrene latex (PSL)

80

2.00

69.13

0.032 ( 0.008

42.7 ( 3.8

0.50 ( 0.13

silver

90

2.00

69.13

0.055 ( 0.012

42.8 ( 5.1

0.61 ( 0.23


100

2.00

69.13

0.042 ( 0.009

36.5 ( 4.0

0.53 ( 0.13


120

2.00

69.13

0.040 ( 0.010

47.8 ( 4.3

0.50 ( 0.15


200

2.00

69.13

0.051 ( 0.011

38.5 ( 4.9

0.40 ( 0.10


250

2.00

69.13

0.041 ( 0.011

42.1 ( 4.0

0.34 ( 0.09


426

2.00

69.13

0.055 ( 0.012

39.8 ( 4.9

0.26 ( 0.07

silicon dioxide

500

2.00

69.13

0.077 ( 0.019

28.1 ( 5.6

0.35 ( 0.08


80

1.00

18.70

0.030 ( 0.007

36.5 ( 3.5

N/A


80

3.20

214.0

0.031 ( 0.007

42.0 ( 3.8

N/A


80

4.00

654.0

0.030 ( 0.006

36.5 ( 3.4

N/A


80

2.00

98.20

0.028 ( 0.006

35.1 ( 3.3

N/A


100

2.00

98.20

0.049 ( 0.015

32.3 ( 4.7

N/A


200

2.00

98.20

0.048 ( 0.013

38.7 ( 4.5

N/A


426

2.00

98.20

0.042 ( 0.011

38.0 ( 4.3

N/A


80

2.00

57.40

0.031 ( 0.006

34.3 ( 3.9

N/A


100

2.00

57.40

0.044 ( 0.010

33.6 ( 3.9

N/A


200

2.00

57.40

0.053 ( 0.014

36.1 ( 4.5

N/A


426

2.00

57.40

0.043 ( 0.009

38.4 ( 3.6

N/A


80

1.00

57.40

0.032 ( 0.007

33.6 ( 3.8

N/A


100

1.00

57.40

0.043 ( 0.011

33.1 ( 4.3

N/A


200

1.00

57.40

0.044 ( 0.010

40.0 ( 4.2

N/A


426

1.00

57.40

0.043 ( 0.010

40.0 ( 4.0

N/A


80

1.00

2.25

0.035 ( 0.009

34.3 ( 3.9

N/A


80

2.54

142.2

0.036 ( 0.009

43.6 ( 3.5

N/A


80

2.00

727.0

0.029 ( 0.007

40.3 ( 3.7

N/A

particle composition

( values represent one standard deviation for parameters from data inversion.


Figure 6. Charge loss during fission from data inversion of TDMA data for electrosprayed 50% methanol, 50% water, 1 mM ammonium acetate.

suspensions on droplet dynamics. As seen from the effective surface tension and percent charge loss data shown in Table 1, none of these parameters had a major influence on charge parameters from data inversion. This supports the notion that, regardless of the initial water and methanol volume fractions, droplets are primarily water with trace amounts of methanol at the end of the evaporation process. The presence of methanol changes the droplet surface tension and field strength for ion emission, affecting both Coulombic fission dynamics and ion emission kinetics. Conclusions TDMA was employed to examine highly charged droplet dynamics in the submicrometer size range and measurements were compared to models of droplet charge loss by Coulombic fission and ion emission. Despite a considerable degree of data variability, a transition from charge loss by Coulombic fission to charge loss by ion emission for water-methanol droplets was found. The observed transition of charge loss in droplets by Coulombic fission to charge loss by single ion emission links studies of micrometer sized droplets18-20 to those of nanometer sized droplets.25,27 Improved understanding of droplet charge loss dynamics may allow for design of electrospay systems better suited for analyte ionization, e.g. use of a solvent from which the field strength necessary for direct emission of an analyte is low. Given the analytical importance of highly charged droplets, further study along these lines is necessary. Acknowledgment. C.J.H. acknowledges support of a National Science Foundation Graduate Research Fellowship (2005-2008). References and Notes (1) Rayleigh, L. Philos. Mag. 1882, 14, 186. (2) Zeleny, J. Phys. ReV. 1914, 3, 69. (3) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass Spectrom. ReV. 1990, 9, 37. (4) Hogan, C. J.; Carroll, J. A.; Rohrs, H. W.; Biswas, P.; Gross, M. L. J. Am. Chem. Soc. 2008, 130, 6926. (5) Hogan, C. J.; Carroll, J. A.; Rohrs, H. W.; Biswas, P.; Gross, M. L. Anal. Chem., in press. (6) Tang, K. Q.; Smith, R. D. Int. J. Mass Spectrom. 1999, 187, 97. (7) Hogan, C. J.; Biswas, P. J. Am. Soc. Mass Spectrom. 2008, 19, 1098. (8) Marginean, I.; Parvin, L.; Heffernan, L.; Vertes, A. Anal. Chem. 2004, 76, 4202.

Hogan et al. (9) Marginean, I.; Znamenskiy, V.; Vertes, A. J. Phys. Chem. B 2006, 110, 6397. (10) Gamero-Castano, M.; Fernandez de la Mora, J. Anal. Chim. Acta 2000, 406, 67. (11) de Juan, L.; Fernandez de la Mora, J. J. Colloid Interface Sci. 1997, 186, 280. (12) Chen, D. R.; Pui, D. Y. H.; Kaufman, S. L. J. Aerosol Sci. 1995, 26, 963. (13) Hogan, C. J.; Yun, K.-M.; Chen, D. R.; Lenggoro, I. W.; Biswas, P.; Okuyama, K. Colloids Surf. A: Physicochem. Eng. Aspects 2007, 311, 67. (14) Collins, R. T.; Harris, M. T.; Basaran, O. A. J. Fluid Mech. 2007, 588, 75. (15) Collins, R. T.; Jones, J. J.; Harris, M. T.; Basaran, O. A. Nature Phys. 2008, 4, 149. (16) Fernandez de la Mora, J.; Loscertales, I. G. J. Fluid Mech. 1994, 260, 155. (17) Gomez, A.; Tang, K. Q. Phys. Fluids 1994, 6, 404. (18) Li, K. Y.; Tu, H. H.; Ray, A. K. Langmuir 2005, 21, 3786. (19) Smith, J. N.; Flagan, R. C.; Beauchamp, J. L. J. Phys. Chem. A 2002, 106, 9957. (20) Taflin, D. C.; Ward, T. L.; Davis, E. J. Langmuir 1989, 5, 376. (21) Gu, W.; Heil, P. E.; Choi, H.; Kim, K. Appl. Phys. Lett. 2007, n/a, 91. (22) Fernandez de la Mora, J. J. Colloid Interface Sci. 1996, 178, 209. (23) Thomson, B. A.; Iribarne, J. V. J. Chem. Phys. 1979, 71, 4451. (24) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287. (25) Ichiki, K.; Consta, S. J. Phys. Chem. B 2006, 110, 19168. (26) Konermann, L. J. Phys. Chem. B 2007, 111, 6534. (27) Znamenskiy, V.; Marginean, I.; Vertes, A. J. Phys. Chem. A 2003, 107, 7406. (28) Fernandez de la Mora, J. Anal. Chim. Acta 2000, 406, 93. (29) Ku, B. K.; Fernandez de la Mora, J. J. Phys. Chem. B 2005, 109, 11173. (30) Ku, B. K.; Fernandez de la Mora, J. J. Phys. Chem. B 2004, 108, 14915. (31) Suh, J.; Han, B.; Okuyama, K.; Choi, M. J. Colloid Interface Sci. 2005, 287, 135. (32) Maze, J. T.; Jones, T. C.; Jarrold, M. F. J. Phys. Chem. A 2006, 110, 12607. (33) Mabbett, S. R.; Zilch, L. W.; Maze, J. T.; Smith, J. W.; Jarrold, M. F. Anal. Chem. 2007, 79, 8431. (34) Cech, N. B.; Enke, C. G. Anal. Chem. 2001, 73, 4632. (35) Benesch, J. L. P.; Robinson, C. V. Curr. Opin. Struct. Biol. 2006, 16, 245. (36) Van Den Heuvel, R. H. H.; Heck, A. J. R. Curr. Opin. Chem. Biol. 2004, 8, 519. (37) Barry, C. R.; Jacobs, H. O. Nano Lett. 2006, 6, 2790. (38) Hogan, C. J.; Biswas, P. Aerosol Sci. Technol. 2008, 42, 75. (39) Kim, H.; Kim, J.; Yang, H.; Suh, J.; Kim, T.; Han, B.; Kim, S.; Kim, D. S.; Pikhitsa, P. V.; Choi, M. Nature Nanotechnol. 2006, 1, 117. (40) Benner, W. H.; Bogan, M. J.; Rohner, U.; Boutet, S.; Woods, B.; Frank, M. J. Aerosol Sci. 2008, 39, 917. (41) Rader, D. J.; McMurry, P. H. J. Aerosol Sci. 1986, 17, 771. (42) Cho, K.; Hogan, C. J.; Biswas, P. J. Nanopart. Res. 2007, 9, 1003. (43) Hogan, C. J.; Li, L.; Chen, D. R.; Biswas, P. J. Aerosol Sci., in press. (44) Jiang, J.; Oberdorster, G.; Biswas, P. J. Nanopart. Res., in press. (45) Rosell-Llompart, J.; Fernandez de la Mora, J. J. Aerosol Sci. 1994, 25, 1093. (46) Chen, D. R.; Pui, D. Y. H.; Hummes, D.; Fissan, H.; Quant, F. R.; Sem, G. J. J. Aerosol Sci. 1998, 29, 497. (47) Kim, S. H.; Woo, K. S.; Liu, B. Y. H.; Zachariah, M. R. J. Colloid Interface Sci. 2005, 282, 46. (48) Knutson, E. O.; Whitby, K. T. J. Aerosol Sci. 1975, 6, 443. (49) Kim, J. H.; Mulholland, G. W.; Kukuck, S. R.; Pui, D. Y. H. J. Res. Natl. Inst. Standards Technol. 2005, 110, 31. (50) Fuchs, N. A. Geofis. Pura Appl. 1963, 51, 185. (51) Li, W.; Li, L.; Chen, D. R. Aerosol Sci. Technol. 2006, 40, 1052. (52) Kousaka, Y.; Okuyama, K.; Adachi, M.; Mimura, T. J. Chem. Eng. Jpn. 1986, 19, 401. (53) Ude, S.; Fernandez de la Mora, J.; Thomson, B. A. J. Am. Chem. Soc. 2004, 126, 12184. (54) Widiyandari, H.; Hogan, C. J.; Yun, K. M.; Iskandar, F.; Biswas, P.; Okuyama, K. Macromolec. Mater. Eng. 2007, 292, 495. (55) Iavarone, A. T.; Williams, E. R. J. Am. Chem. Soc. 2003, 125, 2319. (56) Wyslouzil, B. E.; Wilemski, G.; Strey, R.; Heath, C. H.; Dieregsweiler, U. Phys. Chem. Chem. Phys. 2006, 8, 54. (57) Li, J. S.; Wilemski, G. Phys. Chem. Chem. Phys. 2006, 8, 1266. (58) Nellas, R. B.; Keasler, S. J.; Chen, B. J. Phys. Chem. A 2008, 112, 2930.

JP807765N

Charged Droplet Dynamics in the Submicrometer Size Range - The

Recommend Documents