PARTICLE SENSOR AND SENSING METHOD

Abstract

A particle sensor for measuring size and concentration properties of particles in a gas includes a bipolar diffusion charger configured to charge particles within a received gas sample by the collision of the received particles with and transfer of charge from both positive and negative ions concurrently. At least one electrometer detects the charge of received particles thereby charged. The net, positive, negative or total charge on the bipolarly charged particles has a low sensitivity to variations in the absolute rate of charge generation in the bipolar diffusion charger. A sensor for a ratio of ion charge mobilities in a bipolar diffusion charger employs an ion trap between the bipolar diffusion charger and at least one electrometer.

Claims

1-38. (canceled)

39. A particle sensor comprising an inlet for receiving a gas sample for analysis, a bipolar diffusion charger configured to charge received particles within the received gas sample by the collision of the received particles with and transfer of charge from both positive and negative ions concurrently, and at least one electrometer configured to detect the charge of received particles thereby charged.

40. A particle sensor according to claim 39, wherein the particle sensor comprises an ion trap between the bipolar diffusion charger and at least one said electrometer, wherein the ion trap is configured to remove free positive and negative ions from the mixture of received particles and positive and negative ions formed in the bipolar diffusion charger, before the received particles reach at least one said electrometer.

41. A particle sensor according to claim 40, wherein the ion trap comprises a pair of electrodes spaced apart, which are oppositely charged in use, wherein the current generated from the capture of ions at either of the spaced apart electrodes is measured and the controller processes the current to determine one or more parameters of the charge flux in the bipolar diffusion charger, or to detect an excess of particles, for example to detect smoke in the event of a fire.

42. A particle sensor according claim 40, wherein the ion trap comprises an electrometer and there is at least one further electrometer downstream of the ion trap.

43. A particle sensor according to claim 40, wherein the ion trap is configured to separate the ions from the received particles by diffusion.

44. A particle sensor according to claim 39, wherein the particle sensor further comprises a circuit configured to receive signals from the at least one electrometer and calculate at least one parameter of the concentration and/or size of the received particles in the received gas sample.

45. A particle sensor according to claim 39, wherein charging of the particles reaches a steady state.

46. A particle sensor according to claim 39, wherein at least some of the received particles thereby charged, of the same polarity, are separated from particles of different and/or opposite polarity, before or at the same time as the step of measurement by the at least one electrometer.

47. A particle sensor according to claim 39, wherein the received particles are separated by charge polarity prior to the step of measurement by the at least one electrometer, wherein measurements of the sum of the charge of the separated positively charged particles and the sum of the charge of the separated negatively charged particles are both obtained.

48. A particle sensor according to claim 46, wherein one or more electrometer is configured to both separate at least some of the received particles, of the same polarity, and to measure the charge of the separated particles.

49. A particle sensor according to claim 46, wherein the particles are separated by a potential gradient between electrodes, the potential difference between the electrodes being selected such that the separated particles have a positive charge, and an electrical mobility higher than a threshold, or a negative charge, and an electrical mobility higher than a threshold, and wherein the at least one electrometer measures the separated particles.

50. A particle sensor according to claim 39, wherein the received particles are measured by the at least one electrometer without positive and negatively charged received particles being separated from each other.

51. A method of measuring a parameter of the concentration and/or size of particles in a gas sample for analysis, the method comprising the steps of receiving a gas sample comprising particles, charging received particles in the gas sample by bipolar diffusion charging, thereby charging the received particles by the collision of particles with and the transfer of charge from both positive and negative ions concurrently, and using at least one electrometer to detect the charge of received particles which are thereby charged.

52. A method according to claim 51, wherein as a result of the bipolar diffusion charging, the received particles have a net negative charge or a net positive charge and the net charge is measured by the at least one electrometer.

53. A method according to claim 51, wherein the method comprises separating free positive and negative ions from the mixture of received particles and positive and negative ions formed during bipolar diffusion charging, and wherein the positive and negative ions are removed by an ion trap before the received particles reach at least one said electrometer.

54. A method according to claim 51, wherein the parameters of charge flux of ions within the bipolar diffusion charger is not measured.

55. A method according to claim 51, wherein the received particles are separated from the positive and negative ions by an electrical potential gradient applied between electrodes and the current between the electrodes is not monitored.

56. A method according to claim 51, comprising separating some or all of the positively or negatively charged received particles which have been charged by the bipolar diffusion charger from particles of different and/or opposite polarity, and measuring the charge of the separated positively and/or negatively charged received particles.

57. A method according to claim 51, wherein the received particles are separated by a potential difference applied between two electrodes, and the current at each electrode is measured thereby providing a signal which is a measurement of the charges of the positively charged particles and a signal which is a measurement of the charges of the negatively charged particles.

58. A method according to claim 51, wherein the received particles are separated from the positive and negative ions by an electrical potential gradient applied between electrodes and the current between the electrodes is monitored to measure the rate of ionisation or to detect the presence of a concentration of particles exceeding a threshold or to detect smoke.

59. A method according to claim 51, wherein the charging of the particles by the positive and negative ions does not reach a steady state.

60. A sensor for the ratio of ion charge mobilities in a bipolar diffusion charger, the sensor comprising an inlet for receiving a gas sample for analysis, a bipolar diffusion charger configured to charge particles within the received gas sample by the collision of particles with and transfer of charge from both positive and negative ions concurrently, and at least one electrometer configured to detect the charge of particles thereby charged, the sensor comprising an ion trap between the bipolar diffusion charger and at least one said electrometer, configured to remove free positive and negative ions from the resulting mixture of received particles and positive and negative ions before the received particles reach at least one said electrometer.

61. A method of measuring the ratio of ion charge mobilities in a bipolar diffusion charger, the method comprising the steps of charging particles of known size and/or concentration in a gas sample by bipolar diffusion charging, using a bipolar diffusion charger, thereby charging the particles by the collision of particles with and the transfer of charge from both positive and negative ions concurrently, using at least one electrometer to detect the charge of particles which are thereby charged and processing the measured current to determine the ratio of ion charge mobilities within the bipolar diffusion charger, the method comprising removing free positive and negative ions from the mixture of received particles and positive and negative ions formed during bipolar diffusion charging, using an ion trap, before the received particles reach at least one said electrometer.

Description

DESCRIPTION OF THE DRAWINGS

[0088] An example embodiment of the present invention will now be illustrated with reference to the following Figures in which:

[0089] FIG. 1 is a schematic diagram which is not to scale of a flow-through particle sensor according to a first example of the invention;

[0090] FIG. 2 is a plot of the fraction of particles (y-axis) which have a given charge state (x-axis), expressed as a multiple of the elementary charge, e, following bipolar diffusion charging, both in an experiment (circles) and in theory (squares);

[0091] FIG. 3 is a graph of the mean charge per particle (expressed as a multiple of the elementary charge, e, versus particle diameter, in an experiment using sources of Kr-85 (squares) or Am-241 (triangles), and also as predicted from theory using the Fuchs equilibrium, using Equations 1 to 3 (f.sub.q from Gunn & Wiedensholer using Z.sub.+/Z.sub.−=0.875) (solid line); using Equation 5, for Kr-85 (Gunn approximation using Z.sub.+/Z.sub.−=0.795) (dashed and dotted line); and using Equation 5, for Am-241 (Gunn approximation using Z.sub.+/Z.sub.−=0.889) (dotted line);

[0092] FIG. 4 shows, on the y-axis, the output current from the electrometer of an experimental sensor according to the invention versus, on the x-axis, the output signal from a reference instrument, for measurements taken over a time range and in the same place in an urban environment;

[0093] FIGS. 5A through 5C shows on the x-axis the time of day and on the left hand y-axis measurements in FIG. 5A of particle count N (cm.sup.−3), in FIG. 5B of LDSA (μm.sup.2/cm.sup.3), and in FIG. 5C of aerosol current, I (fA), for measurements taken over a time range and in the same place in an urban environment using in FIG. 5A a condensation particle counter (N), FIG. 5B a reference device (LDSA), and in FIG. 5C a device according to the present invention (i);

[0094] FIG. 6 is a schematic diagram of a second embodiment of the invention; and

[0095] FIG. 7 is a graph of the calculated expected absolute values of (i) the mean charge per particle for positively charged particles (82), (ii) the mean charge per particle for negatively charged particles (80), (iii) the difference between (i) and (ii) (the net mean charge per particle) (86) and the sum of (i) and (ii) (the total mean charge per particle) (84) in particles charged by a bipolar diffusion charger at standard conditions, versus particle diameter.

[0096] FIG. 8 is a graph of the calculated expected absolute values of (i) the current for positively charged particles (82′), (ii) the current for negatively charged particles (80′), (iii) the difference between (i) and (ii) (the net aerosol current) (86′) and the sum of (i) and (ii) (the total aerosol current) (84′) in particles charged by a bipolar diffusion charger versus particle diameter, assuming 10,000 particles per cubic centimetre of gas at standard conditions with a sample flow-rate of 0.3 L/min.

DETAILED DESCRIPTION OF AN EXAMPLE EMBODIMENT

First Example—Measuring Net Aerosol Current

[0097] With reference to FIG. 1, a first example of particle sensor 1 has a body 2 defining an inlet 4 and an outlet 6. A fan 8 is configured to draw an air sample 10 along an air flow path in use, through a chamber which functions as a bipolar diffusion charger 12, a conduit 14, which extends from the bipolar diffusion charger to an ion trap 16, and then to a particle detection chamber 18 which comprises a Faraday cup 20, functioning as an electrometer.

[0098] The invention is for analysing particles 50, 50a, 50b within an air sample 10 (the received particles). Particles are shown schematically and not to scale. In FIG. 1, neutrally charged particles 50 are shown with an open circle, a diagonally striped circle indicates a positively charged particle 50b and a circle with horizontal lines indicates a negatively charged particle 50c. Although the Figures show only neutrally charged particles at the inlet, in practice ambient air includes some charged particles and/or free gas ions. Therefore some received particles may already be charged, with a charge distribution which depends on, inter alia, their source and age. However, the equilibrium processes of bipolar diffusion charging will mean that variations in the charge distribution in received particles will, unless highly charged, have negligible effect.

[0099] The bipolar diffusion charger 12 includes an ionizing radiation source 22, which in an example is formed by three units of Americium-241 (alpha decay, 432.2 year half life), spaced apart along a flow chamber, and with a combined activity of under 111 kBq, and in another example is formed by Krypton-85 (beta decay, 10.76 years half life), 370 MBq.

[0100] The ion trap takes the form of a flow-through electrostatic precipitator which has opposed and oppositely charged electrodes 24, 26 and a circuit configured to maintain a potential difference between these electrodes in use. The electrometer comprises a current sensor 30, having an output 32 which is connected to an input of a controller 34 which in turn has a signal output 36 extending to an output interface 38. The controller typically comprises a processor (such as a microprocessor or microcontroller) executing stored code although its function may be implemented in whole or part with discrete electronic components.

[0101] In use, the air sample comprising uncharged particles 50 which are to be analysed, is drawn through the sensor by the action of the fan, into the bipolar diffusion charger 12. Ionizing radiation from the radiation source 22 ionises gas molecules to form a mixture of positively charged gas ions 60a and negatively charged gas ions 60b. Overall, these gas ions within the bipolar diffusion charger have a net zero charge. No electrical potential gradient is applied within the bipolar diffusion charger and the ions move by diffusion and by virtue of electrostatic forces (between charged particle and ions rather than by virtue of an externally applied electric field). The particles in the flowing gas (the received particles) sample collide with the positively and negatively charged ions and become charged so that there are concurrently formed both positively and negatively charged particles (50a, 50b). Some particles continue to have, or regain, a net charge of zero. As particles flow through the conduit, the charge distribution continues to tend to an equilibrium. As we will explain further below, the net result of this bipolar diffusion charging is that the individual particles have a statistical distribution of charges but with an overall net negative charge. Bipolar diffusion charging uses a mixture of both positive and negative ions and the diffusion of these ions to charge particles, obtaining a mixture of positive and negative ions concurrently.

[0102] Within the ion trap, the potential difference applied between the electrodes 24, 26 is selected to be sufficient to cause the free gas ions 60a, 60b which remain in the flowing gas sample to be drawn to the electrodes and therefore separated from the particles 50, 50a, 50b. This is readily achieved by selection of an appropriate potential gradient as the particles 50, 50a, 50b typically have a substantially lower mobility than the gas ions 60a, 60b, usually by a factor of at least 100 for particles with a diameter of 10 nm (and a greater factor for larger particles). The particles 50, 50a, 50b, without the gas ions, continue to flow towards the detector 18.

[0103] Within the detector 18, a Faraday cup electrometer 20 collects all of the particles, including the neutral particles, using a high-efficiency particulate air (HEPA) filter. The flux of charged particles into the electrometer induces an image current on the Faraday cup, which is measured by the ammeter 30. Air, which is generally free of particles, passes out of the outlet 6.

[0104] The controller 34 processes the measured current and the air speed to calculate a value of the first moment of the particle diameter distribution per unit volume, or a related property such as lung deposited surface area per unit volume, which it outputs as a digital or analogue signal throughout output 38. If the speed of rotation of the fan 8 is sufficiently controlled, the speed of air flow may be known, and a constant scaling factor may be employed. Alternatively, an anemometer or other air flow sensor may be used to measure air flow velocity. In some embodiments, no fan is present, either because air is drawn or pushed through the filter at a known speed, or it is judged sufficient to measure the speed of air flow.

[0105] The invention exploits several properties of the equilibrium charge distribution arising from bipolar diffusion charging. By way of explanation, at bipolar charge equilibrium, the fraction of total particles (f.sub.q) at each charge state (q) as a function of diameter greater than 50 nm is estimated as follows by Gunn R. and Woessner R. (Measurement of the systematic electrification of aerosols, Journal of Colloid Science, 1956, 11, 254-259):

[00001]$\begin{array}{cc}{f}_q\left(d\right)=\frac{e}{\sqrt{2\pi G}}\exp \left(-\frac{{\left(q-{\mathrm{Ge}}^{-2}\ln \left(Z_{+}/Z_{-}\right)\right)}^2}{2{\mathrm{Ge}}^{-2}}\right)& \left(1\right)\end{array}$$\mathrm{where}$$\begin{array}{cc}G=2{\mathrm{\pi \epsilon }}_0\mathrm{dkT}& \left(2\right)\end{array}$

for a given absolute temperature (T), positive to negative ion mobility ratio (Z.sub.+/Z.sub.−), and the constants of electron charge (e), vacuum permittivity (ε.sub.0) and Boltzmann's constant (k). The mean charge per particle (q.sub.d) for a given particle diameter (d) is given as the sum:

[00002]$\begin{array}{cc}{\stackrel{\_}{q}}_d=\underset{q}{.Math.}{\mathrm{qf}}_q\left(d\right)& \left(3\right)\end{array}$

and the corresponding current as a function of diameter (i.sub.d) is:

[00003]$\begin{array}{cc}{i}_d={\mathrm{QeN}}_d{\stackrel{\_}{q}}_d& \left(4\right)\end{array}$

[0106] Where Q is the volumetric flow rate of the gas containing the charged aerosol and N.sub.d is the number concentration of particles of diameter, d. The summation over the discrete charge states of Equation 4 requires a numerical calculation which can be analytically approximated as shown in Equation 5.

[00004]$\begin{array}{cc}{\stackrel{\_}{q}}_d=\int {\mathrm{qf}}_q\left(d\right)\mathrm{dq},& \left(5\right)\end{array}$${\stackrel{\_}{q}}_d=\frac{i_d}{N_d\mathrm{Qe}}=\frac{2{\mathrm{\pi ϵ}}_0\mathrm{kT}}{e^2}\ln \left(\frac{Z_{+}}{Z_{-}}\right)d.$

[0107] At standard conditions the largest difference between the integral approximation (Eq. 5) and the discrete charge summation (Eq. 3) is 0.5%. This disagreement occurs for particles of 50 nm, which is about the minimum size for which these particular equations are valid, although the sensor is useful for measuring smaller particles too. FIG. 2 shows the high correlation between theory and measured charge states in an example.

[0108] Equation 5 states that the mean charge per particle (q.sub.d) and corresponding net current (i.sub.d) for a given particle diameter (d) is a simple linear function of the particle diameter (d), the gas temperature (T), and the natural log of the ratio of positive to negative ion mobilities (Z.sub.+/Z.sub.−). The mean charge (q) carried by an aerosol with a polydispersed distribution of particle sizes is given by the integral sum of all electrical charges over all diameters, given by the distribution f.sub.d:

[00005]$\begin{array}{cc}\stackrel{\_}{q}={\int }_0^{\infty }{\stackrel{\_}{q}}_d\left(d\right)f_d\left(d\right)\mathrm{dd}.& \left(6\right)\end{array}$

[0109] The corresponding net aerosol current can be derived therefrom as:

[00006]$\begin{array}{cc}i=\left[\mathrm{Qe}\frac{2{\mathrm{\pi \epsilon }}_0\mathrm{kT}}{e^2}\ln \left(\frac{Z_{+}}{Z_{-}}\right)\right]N\stackrel{\_}{d}.& \left(7\right)\end{array}$

where d is the mean particle diameter of the polydispersed aerosol. Note q is the mean charge per particle over a polydisperse distribution whereas q.sub.d (of Eq. 5) is the mean charge per particle at one particle size.

[0110] From Equation 7, it is clear that the only knowledge required to determine Nd is the measured current, the gas flow rate and temperature, and the ratio of ion mobilities. The former can be measured accurately, while the latter can be determined for common mixtures. The net current detected at the Faraday cup electrometer will therefore be approximately proportional to Nd for a range of particle diameters and accordingly the current may be used to estimate the LDSA concentration of aerosols within the received sample.

[0111] The particle sensor may therefore also be useful for measuring the net ratio of ion mobilities in a bipolar charger. For particles of a known size and concentration, the net ratio of ion mobilities in the bipolar charger can be calculated from the measured current, taking into account known or measured gas flow rate and known or measured temperature, using Equation 7. In this application, particles with a known size distribution and so known Nd are passed through a bipolar diffusion charger at a known or measured volumetric flow rate, at a known or measured temperature (and possibly humidity) and the measured current is processed to calculate the net ratio of ion mobilities within the bipolar diffusion charger. The resulting ratio is useful in a number of applications, for example it may be used to correct equilibrium charge states for use in the SMPS inversion algorithm for multiple-charge correction an SMPS devices. The ratio may be calculated to determine the effects of changes in gas composition on the ratio of ion mobilities.

[0112] FIG. 3 demonstrates that in a practical embodiment, the mean charge per particle obtained using the radiation sources described above is a linear function of particle diameter, as predicted from theory, from about 50 nm to about 1000 nm. The relationships remain reasonably accurate below and above this range. Eventually, for sufficiently large particles, it is overtaken by other effects.

[0113] Spherical particles (Bis(2-ethylhexyl) sebacate; DOS) were atomized using a nebulizer with HEPA-filtered, compressed air. A sample from the main flow of spherical particles was diluted using a disk diluter to provide a range of particle number concentrations (dilution ratios between 10 and 150), while the remaining aerosol was vented. A custom-built electrostatic precipitator (ESP) removed any particles charged during atomization or dilution. The diluted aerosol sample was then classified by an aerodynamic aerosol classifier to generate an aerodynamically monodispersed source. Since the DOS formed nano-droplets (i.e. spherical particles of known density), the AAC classified particles were also monodispersed in particle geometric diameter which was calculated at each AAC setpoint. It should be noted that for spherical particles this geometric diameter is equivalent to the particle mobility diameter, a parameter commonly used by others to validate bipolar and unipolar diffusion charging theory (Gopalakrishnan, R.; Thajudeen, T.; Ouyang, H.; Hogan Jr, C. J. The unipolar diffusion charging of arbitrary shaped aerosol particles. Journal of Aerosol Science 2013, 64, 60-80; and Gopalakrishnan, R.; McMurry, P. H.; Hogan Jr, C. J. The bipolar diffusion charging of nanoparticles: A review and development of approaches for non-spherical particles. Aerosol Science and Technology 2015, 49, 1181-1194.)

[0114] The neutral, monodispersed particles were sampled by a condensation particle counter (CPC) to measure particle number concentration (N) in parallel with the proof-of-concept measurement device. The disk diluter and AAC controlled the particle number concentration and size, respectively. The mean charge per particle can be determined based on the electrical current measured by the aerosol electrometer operating with flow rate (Q) and total particle number concentration (N) using Equation 5. This mean charge per particle (q) is compared against the one predicted by theory using Equation 5 with the total particle number concentration (N) measured by the CPC and the mean particle diameter selected by the AAC (d).

[0115] FIG. 4 shows the current obtained from the detector of a prototype device according to the present invention in comparison with a Naneos Partector reference instrument which is considerably more expensive than a sensor as described herein would cost (Naneos and Partector are trademarks of Naneos Particle Solutions GmbH, Windisch, Switzerland). FIGS. 5A, 5B and 5C show the output signal for measurements made near an urban road over time in FIG. 5A, particle count information from a condensation particle counter (number concentration in units of particles per cubic centimetre), FIG. 5B, the Naneos Partector (LDSA in units micrograms per cubic meter) and FIG. 5C, a prototype device for the present embodiment (current in units of femtoAmps, which is proportional to LDSA).

[0116] It can be seen that the detector according to the invention has an output which correlates closely with the reference device but in a simple product which can be manufactured at low cost. The output current is proportional to the first moment of particle size distribution and can therefore be used to estimate lung deposited surface area, whereas the reference device measures an output signal approximately proportional to the first moment of particle size distribution, subject to more environmental variables. The device can be substantially miniaturised, with the practical size of electronics likely to be the limiting size consideration. The radiation sources are comparable to or may be the same as those used in everyday smoke alarms and do not require special handling procedures for consumers. In contrast to, for example ionisers based on corona discharges, no ozone or other environmentally damaging gases are produced.

[0117] The invention may use any bipolar ion source which produces an equilibrium charge distribution on particles. Embodiments for use in environments with high particle concentrations, or where high sample flow rates are required, may require ion sources which produce a greater excess of ions. Nevertheless, it is not essential that the bipolar charging of the particles actually reaches a steady state. Where insufficient ions are produced, or the particle residence time is too short (e.g. due to a high air speed), a steady state (equilibrium) might not be reached, but there will be a functional relationship between particle charge and mean diameter and concentration.

[0118] An important feature of the invention is that the net charge on particles is relatively stable and, provided that an excess of ions is generated, is insensitive to variations in the rate of ion generation and the residence time of particles exposed to ions (in contrast to unipolar ion chargers, for example), provided they are above a minimum threshold. Accordingly, the embodiment described above with reference to FIG. 1 does not require to include any means of measuring or controlling the rate of ion generation.

[0119] Nevertheless, in some embodiments, the current which flows between the electrodes, 24, 26, could be measured and used as an indicator as to when a steady state is not being reached. For example, in this case, the current generated from the capture of ions at either of the opposed electrodes would be lower than anticipated. This could be used to determine that the ion source is no longer functioning effectively. However, a decrease in the ion current over a relatively short period of time may indicate that there is an excessively high concentration of particles present. This could be useful to indicate that the device may not provide an accurate reading, or to provide a smoke detector function or an alarm for a high concentration of aerosols.

[0120] It is also possible to determine when the charge equilibrium is not being adequately achieved from the current at the one or more electrometers. If the first moment of particle size distribution (Nd) exceeds a threshold the total current will level out as there will be insufficient ions to enable equilibration for the particles which are present. Accordingly, an alert can be generated responsive to the current at the one or more electrometers exceeding a predetermined threshold.

[0121] Where required, the sensitivity of the device can be improved by employing modulation. For example, the ion trap might be switched on and off according to a predetermined pattern, e.g. square wave, and the resulting modulation detected in the current measured at the electrometer. It would also be possible to modulate the ion source, for example with a shutter or gate between the radioactive element and the bulk of the volume of the bipolar diffusion chamber.

Second Example—Separation of Positively and/or Negatively Charged Particles

[0122] In a second example embodiment, illustrated with reference to FIG. 6, particles which have been charged by bipolar diffusion charging are separated by polarity, i.e. the positive and negative particles are separated from each other. It is not critical whether the neutrally charged particles are, or are not, separated from other particles. Numbered features in FIG. 6 correspond to correspondingly numbered features of FIG. 1. The bipolar diffusion charger 12 and ion trap 16 are as before. However, particle detection chamber 18 and Faraday cup electrometer 20 are replaced with a particle separation and detection chamber 70, which is downstream of the ion trap and which comprise a positively charged electrode 72 and an opposite negatively charged electrode 74. First and second ammeters 76 and 78 measure the current in each electrode in use. The potentials of the positively charged and negatively charged electrodes are selected to cause positively charged particles 50b to be captured by the negatively charged electrode 74, and their charge detected by the second ammeter, giving a current i.sub.+, while the negatively charged particles 50a are captured by the positively charged electrode 72 and their charge detected by the first ammeter, giving a current L. Neutrally charged particles pass through the chamber and the outlet 6. As apparent from FIG. 2, the positively charged particle current, i.sub.+, comprises contributions from particles having each of a plurality of different positive charges, and the negatively charged particle current, i.sub.−, comprises contributions from particles having each of a plurality of different negative charges.

[0123] The controller 34 processes i.sub.+ and i.sub.− to determine size parameters of the particles, including the LDSA and/or first moment of the size distribution (Nd) of the particles, with reference to calculated or calibrated equivalence data representing the relationship between these currents, and number concentration and size parameters. The controller may calculate abs [i.sub.+]+abs[i.sub.−] (abs referring to the absolute value of), referred to herein as the total aerosol current and/or abs [i.sub.+]−abs [i.sub.−], the net aerosol current, which corresponds to the current which would be measured using the first example sensor.

[0124] Following a similar derivation to predict the net mean charge per particle (Eqn 5) of the first embodiment, the mean charge per particle for positively (q.sub.d.sub.+) or negatively (q.sub.d.sub.−) charged particles of given diameter (d) separated and measured by the second embodiment can be estimated using:

[00007]$\begin{array}{cc}{\overline{q}}_{d_{+}}={\int }_0^{\infty }q.Math.f_q\mathrm{dq}=\frac{\sqrt{G}}{e\sqrt{2\pi }}\exp \left(-C^2\right)+\frac{G}{2e^2}.Math.\ln \left(\frac{Z_{+}}{Z_{-}}\right)\mathrm{erfc}\left[-C\right]& \left(7\right)\end{array}$$\begin{array}{cc}{\overline{q}}_{d_{-}}={\int }_{-\infty }^{0}q.Math.f_q\mathrm{dq}=-\frac{\sqrt{G}}{e\sqrt{2\pi }}\exp \left(-C^2\right)-\frac{G}{2e^2}.Math.\ln \left(\frac{Z_{+}}{Z_{-}}\right)\left(-2+\mathrm{erfc}\left[-C\right]\right)& \left(8\right)\end{array}$

where G is defined by Eqn 2 and C is:

[00008]$\begin{array}{cc}C=\frac{\sqrt{G}}{e\sqrt{2}}\ln \left(\frac{Z_{+}}{Z_{-}}\right).& \left(9\right)\end{array}$

[0125] These mean charge per particle based on particle polarity correspond to the following aerosol currents as a function of particle diameter:

[00009]$\begin{array}{cc}{i}_{d+}={\mathrm{QeN}}_d{\overline{q}}_{d+}={\mathrm{QeN}}_d(\frac{\sqrt{G}}{e\sqrt{2\pi }}\exp \left(-C^2\right)+\frac{G}{2e^2}.Math.\ln \left(\frac{Z_{+}}{Z_{-}}\right)\mathrm{erfc}\left[-C\right]& \left(10\right)\end{array}$$\begin{array}{cc}{i}_{d-}={\mathrm{QeN}}_d{\overline{q}}_{d-}={\mathrm{QeN}}_d(-\frac{\sqrt{G}}{e\sqrt{2\pi }}\exp \left(-C^2\right)-\frac{G}{2e^2}.Math.\ln \left(\frac{Z_{+}}{Z_{-}}\right)\left(-2+\mathrm{erfc}\left[-C\right]\right)& \left(11\right)\end{array}$

[0126] FIG. 7 shows the predicted magnitude of the mean charge per particle from different components of the aerosol sample assuming gas at standard conditions as a function of different particle diameters; specifically the magnitude of the mean charge per particle for positively charged particles, q.sub.+ 82, the magnitude of the mean charge per particle for negatively charged particles, q.sub.− 80, the sum of these, the total mean charge per particle, abs [q.sub.+]+abs [q.sub.−], 84, and the magnitude of the difference between the positively charged particle and the negatively charged particles, the net mean charge per particle, abs [q.sub.+]−abs [q.sub.−], 86. In each case, dashed lines are numerical solutions predicted by Gunn & Wiedensohler, and the dotted lines are an analytical solutions derived from Gunn to approximate the numeric solutions. This figure is an expansion of FIG. 3 which only shows the net mean charge per particle.

[0127] FIG. 8 shows the predicted magnitude of the currents from different components of the aerosol sample assuming 10,000 particles per cubic centimetre of gas at standard conditions with a sample flow-rate of 0.3 L/min as a function of different particle diameters; specifically the positively charged particle current, i.sub.+, 82′, the negatively charged particle current, .sub.−, 80′, the sum of these, the total aerosol current, abs [i.sub.+]+abs [i.sub.−], 84′ and the magnitude of the difference between the positively charged particle current and the negatively charged particle current, the net aerosol current, abs [i.sub.+]-abs [i.sub.−], 86′. In each case, dashed lines are numerical solutions predicted by Gunn & Wiedensohler, and the dotted lines Whereas FIG. 7 shows the absolute mean charge per particle (abs[q]), FIG. 8 shows the corresponding aerosol current (i) assuming a particle concentration and sample flow-rate.

[0128] It can be seen that the positively charged particle current, the negatively charged particle current, and the total aerosol current are each substantially larger than the net aerosol current, abs [i.sub.+]−abs [i.sub.−], particularly at lower particle diameters. We calculate that the ratio of total aerosol current to net aerosol current would be about 1900% for 2 nm diameter particles. This increase in measurable current may greatly improve measurement accuracy, particularly at low particle sizes where currents are smallest and prone to error.

[0129] Still further, as two independent current measurements are made, the controller 34 can independently calculate the number density of particles (N) and their mean diameter (d), rather than only the first moment of the particle size distribution. The independent currents may have different properties, for example the positively charged particle current, i.sub.+, is much less sensitive to particle size than the negatively charged particle current, i.sub.+. Thus the positively charged particle current, i.sub.+, alone might give a useful indication of particle concentration within a range of uncertainty. This is relevant for example to periodic technical inspection requirements in Germany which allow a relatively high uncertainty for number concentration measurements.

[0130] In the above example, all positively charged particles are separated from all negatively charged particles. However, it would suffice to separate positively charged particles having at least a threshold electrical mobility or negatively charged particles having at least a threshold electrical mobility from particles of the opposite charge (and typically also from the neutrally charged particles), for example by applying an offset to the potential of both electrodes 72, 74, and in some circumstances this would provide additional information.

[0131] It is alternatively possible to measure the positive and negative particle currents i.sub.+ and i.sub.− using apparatus like that of FIG. 1 further including a trap which can selectively remove positive or negative particles, enabling the measurement of the charge of the remaining particles, which is modulated in use. Alternatively, positively and negatively charged particles may be separated using an electromagnetic field and detected by separate electrometers 20.

[0132] In the examples above, ions are formed in air molecules with properties and in a ratio which depends on the type and energy of radioactive decay as well as the state and composition of the carrier gas, which affects the net particle charge for a given diameter. In some embodiments, the properties of the ions in the bipolar diffusion charger are manipulated. One way in which this can be achieved is to introduce new components to the received gas, either before the sample gas reaches the bipolar diffusion chamber, or within the bipolar diffusion charger. It is also possible to vary properties of the charger or the ion source. Changes in the difference between ion mobility for positive and negative charges affect the equilibrium charge distribution of the aerosol, and so one or more components of the gas or properties of the charger or ion source may be selected to increase the signal strength of the device.

[0133] In some embodiments, the bipolar diffusion charger comprises a salt such as NaCl or other easily ionisable chemicals (e.g. siloxanes from silicone), which would be preferentially ionised. The presence of these chemicals changes the composition of the ions which are generated, especially the relative mobility of positive and negative ions, and therefore the charge distribution. In some embodiments there is provided a temperature regulator (e.g. heater or cooler) to control the temperature of the ions within the bipolar diffusion charger.

[0134] The ion trap may in some embodiments not use an electrical potential gradient, but instead ions may be trapped at walls via Brownian diffusion or electrostatic forces. The ion trap may be an ion selective membrane. These embodiments provide a simpler device.

[0135] Further variations and modifications may be made within the scope of the invention herein disclosed.