Field of invention

The present invention relates to an apparatus, i.e. a sensor, for measuring particle mass concentration according to the preamble of claim 1 and specifically to an apparatus allowing the removal of ultrafine particles from the particle stream. The present invention further relates to the use of a particle sensor for measuring particle mass concentration according to the preamble of claims 8, 12 and 14. The present invention further relates to a process for measuring particle mass concentration according to the preamble of claim 18. Background of the invention

There is a constant increase in the demand for real-time particle control. Especially the real-time exhaust control of combustion engines, such as vehicles, requires reliable and non-expensive particle monitoring. Requirement for particle control exists also e.g. in indoor air quality monitoring or with air traffic safety. The particle amount is in most cases expresses as particle mass concentration, in mg/m ³ or equivalent.

Various particle measurement devices are based on electrically charging particles and measuring the electrical current carried by such charged particles. One such prior art method and apparatus for measuring fine particles is described in document WO2009109688 Al, Pegasor Oy, 11.11.2009. In this prior art method clean, essentially particle free, gas is supplied into the apparatus and directed as a main flow via an inlet chamber to an ejector provided inside the apparatus. The clean gas is further ionized before and during supplying it into the inlet chamber. The ionized clean gas may be preferably fed to the ejector at a sonic or close to sonic speed. The ionizing of the clean gas may be carried out for example using a corona charger. The inlet chamber is further provided with a sample inlet arranged in fluid communication with a channel or a space comprising aerosol having fine particles. The clean gas flow and the ejector together cause suction to the sample inlet such that a sample aerosol flow is formed from the duct or the space to the inlet chamber. The sample aerosol flow is thus provided as a side flow to the ejector. The ionized clean gas charges the particles. The charged particles may be further conducted back to the duct or space containing the aerosol. The fine particles of the aerosol sample are thus monitored by monitoring the electrical charge carried by the electrically charged particles. Free ions may removed further be removed using an ion trap. A major problem in any particle measurement device which is based on electrically charging particles and measuring the electrical current carried by such charged particles is the conversion of the measured electrical current to actual mass concentration. The conversion factor may be simply determined by calibrating a particle measurement device against a reference method, typically against gravimetric method, which accurately determines mass concentration. However, such calibration may change due to changes in the shape of particle size distribution curve, mean particle diameter, width of a lognormal particle size distribution curve, particle shape (usually expresses with fractal parameters) or particle density.

United States Patent US 7,812,306 B2, TSI, Incorporated, 12.10.2010, describes an instrument for non- invasively measuring nanoparticle exposure includes a corona discharge element generating ions to effect unipolar diffusion charging of an aerosol, followed by an ion trap for removing excess ions and a portion of the charged particles with electrical mobilities above a threshold. Downstream, an electrically conductive HEPA filter or other collecting element accumulates the charged particles and provides the resultant current to an electrometer amplifier. The instrument is tunable to alter the electrometer amplifier output toward closer correspondence with a selected function describing particle behavior, e.g. nanoparticle deposition in a selected region of the respiratory system. Tuning entails adjusting voltages applied to one or more of the ion trap, the corona discharge element and the collecting element. Alternatively, tuning involves adjusting the aerosol flow rate, either directly or in comparison to the flow rate of a gas conducting the ions toward merger with the aerosol. The publication is focused on the measurement of particle concentrations in terms of surface area, as such accumulated or aggregate surface area are expected to provide more useful assessments of health risks due to nanoparticle exposure. The publication actually teaches that mass concentration measurements are not useful in indicating health effects and thus would not motivate person seeking for a solution on converting measured electrical current into mass concentration to examine the technique described in the publication. United States Patent US 8,122,711 B2, Robert Bosch GmbH, 28.2.2012, concerns a procedure to ascertain a concentration of sooty particles in an exhaust gas system of an internal combustion engine or a depletion of an emission control system of the internal combustion engine due to the loading of sooty particles, whereby the sooty particle concentration in the exhaust gas system is determined by means of a collecting particle sensor, which emits a sensor signal and whereby the depletion of the emission control system due to the loading of sooty particles is determined from the sooty particle concentration. The sensor signal is corrected by means of predetermined corrections with regard to a sensor temperature and/or an exhaust gas temperature and/or a flow velocity of the exhaust gas and/or a voltage applied at the particle sensor. Transverse sensibilities of the particle sensor can thereby be taken into account during the evaluation; and the determination of the accumulated loading of sooty particles and the determination of the sooty particle concentration in the exhaust gas system are improved. In the process, the sensor temperature enters into the correction to the extent that a temperature dependence of the electrical resistance of the loading of sooty particles is determined in a preparation phase and can be taken into account during the evaluation of the sensor signal. Although the procedure improves the mass

concentration measurement, it involves extra components and is thus clumsy and costly.

Thus the particle sensors of the prior art possess the technical problem of the electrical current signal vs. mass concentration ratio being sensitive to external conditions. There is need for a sensor which can measure or monitor particle mass concentration even when the particle mean diameter is changing. Brief description of the invention

The object of the present invention is to provide an apparatus so as to overcome or at least alleviate the prior art disadvantages. The objects of the present invention are achieved with an apparatus according to the characterizing portion of claim 1. The objects of the present invention are also achieved with a process according to the characterizing portion of claim 14. The objects of the use of the invented apparatus are achieved according to the characterizing portion of claims 4 and 8.

The preferred embodiments of the invention are disclosed in the dependent claims.

The inventor has surprisingly found a process which will solve the prior art problems described above. The invented process is based on decreasing the current signal generated by charged particles by trapping a fraction of the charged particles. Such a process leads to significant improvement in reducing the measurement error generated by changing count median diameter (CMD) of particles under measurement.

The electrical current carried by charged particles depends on the particle size distribution of the measured particles. Lognormal particle size distribution has been found to apply to most single-source aerosols such as e.g. combustion engine or stack emission aerosols. The lognormal distribution is used extensively for aerosol size distributions because it fits the observed size distributions reasonably well, which is obvious as such for a person skilled in the art. The lognormal distribution can be characterized with the median particle diameter, which is different for number concentration and mass concentration, the median particle diameter for number concentration being smaller than the median particle diameter for mass concentration. Particle mass depends also on particle density and particle shape, which is frequently characterized by fractal dimensions. The fractal dimensions are well known for a person skilled in the art and are described e.g. in William C. Hinds, Aerosol Technology - Properties, Behavior, and Measurement of Airborne Particles, 2 ^nd edition, Jonhn Wiley & Sons, 1999, pages 408-412 and in Maricq and Wu, The effective density and fractal dimensions of soot particles from premixed flames and motor vehicle exhaust, Aerosol Science 35 (2004), pp. 1251-1274. With single-source aerosols the changes in the conditions affecting particle formation mainly change the median diameter, other parameters affecting particle charging remaining essentially persistent.

The invented process for particle mass concentration measurement comprises guiding sample flow Q comprising particles through a passage, electrically charging particles and measuring the electrical current carried by said charged particles. The invented process further comprises trapping essentially all free ions and a particles (P) with diameter smaller than cut-off diameter D _c _o by an electrical field with trap voltage V, adjusting the trap voltage V in such a way that the measured electrical current I is smaller than the electrical current l ₀ measured when the trap voltage V is set to V ₀ trapping essentially all free ions only and converting the electrical current signal to particle mass concentration value.

Particle measurement, which includes particle charging and measurement of the electrical charge carried by charged particles requires that the free ions, i.e. ions which are not attached to particles, are removed. The design of an ion trap used to remove the free ions can be carried out either by calculations or by experiments. Ion trapping is based on the electrical mobility of charged particles, Z _p = n e C _c / 3 π η D _p , where Z _p = Electrical mobility, n = Number of charges on the particle, e = Elementary charge, C _c =

Cunningham slip correction, η = Dynamic viscosity of air and D _p = Radius of particle. As essentially all free ions need to be removed, the ion trap is typically designed to trap also ultrafine charged particles, which - due to their low concentration and mass - do not essentially affect the number or mass concentration measurement. A typical design of an electrical trap would have a particle cut-off diameter of 4 nm, i.e. it removes all charged particles having a diameter of 4 nm or less.

The conversion factor which is used to convert the measured electrical current, usually expressed in fA to mass concentration value, usually expressed in mg/m ³ , depends on the mechanism of particle formation. Within similar single-source particles, the conversion factor depends mainly on the median particle diameter. Thus the conversion factor should be changed as the median particle diameter changes to acquire a reliable current-to-mass concentration conversion. However, as the changes in the median particle diameter are not known, such conversion factor adjustment cannot be made. For a well-known particle source, such as for particle formation within combustion engines, the changes in the median diameter as a function of a second-grade parameter, such as e.g. combustion engine torque could be determined. Information in the change of the second-grade parameter (such as engine torque) could then be used to adjust the conversion factor. Such operations are, however, clumsy and expensive. Surprisingly the inventor has found that the sensitivity of the correlation factor to the changes of the median diameter can be significantly reduced by increasing the trap voltage in such a way that the ion trap normally used to trap essentially all free ions only (i.e. particle cut-off diameter around 4 nm) is used to trap charged particles so that the particle cut-off diameter of the voltage trap is significantly higher than 4 nm. In such case the measured electrical current I for a sample flow Q with particles is lower than the electrical current l ₀ measured for the same sample flow but using a trap voltage which removes essentially all free ions. The l/l ₀ ratio is essentially constant, and the effect of the lower current value into the mass/current conversion factor can be determined either experimentally or by calculations.

Surprisingly the trap voltage V is advantageously adjusted in such a way that the measured electrical current I is preferably 20-95% smaller and even more preferably 65-85% smaller than the electrical current l ₀ measured when the trap voltage V is set to V ₀ trapping essentially all free ions only, i.e. typically trapping essentially all charged particles with less than 4 nm diameter.

It is to be noted that as the electrical current is mainly carried by small particles whose count is larger than the count of large particles where most of the particle mass is, the influence of reducing the current signal does not as much affect the amount of particle mass flowing through the measurement passage. Thus the trap voltage can be preferably adjusted to a value where the trap cut-off diameter D _c . ₀ is essentially equal to the count median diameter (CMD) of particles in sample flow Q. Essentially in this case means e.g. that the cut-off diameter D _c _ ₀ is within -50% - +100% of the CMD of particles in sample flow (Q).

The CMD of the particles in the sample flow can be determined by various measurement techniques, e.g. by using Electrical Low Pressure Impactor (ELPI) by Dekati Oy, Finland. Typically the CMD of e.g. diesel emissions depends on the type of the diesel motor, the fuel and the sampling method, such as described in e.g. Wong, CP.; Chan, T.L.; Leung, C.W., Characterization of diesel exhaust particle number and size distributions using mini-dilution tunnel and ejector-diluter measurement techniques, Atmospheric Environment, Volume 37, Number 31, October 2003 , pp. 4435-4446(12), the contents of which are enclosed herewith for reference purposes, but stays within -50% - +100% of the nominal CMD during the use of the specific motor and fuel. Thus the nominal CMD values may be determined for combinations of interest and they can be used as the set cut-off diameter D _c . ₀ of the present invention. Reasonably accurate CMD's for different combustion processes can also be found in the literature.

The CMD used within the current invention is thus a nominal CMD, CMD _nom , for a specific emission source and sampling conditions, determined either by well-known measurement methods, such as the ones described in the article of Wong et al, or otherwise known with a reasonably accuracy e.g. from reference literature.

In one embodiment of the present invention the trap voltage is adjusted during the measurement. This may be done when the process producing particles operates in steady state, e.g. when a combustion engine operates with essentially constant torque. First the trap voltage is set to a value, which ensures the removal of essentially all free ions, e.g. the trap voltage is set to a value which corresponds to particle cutoff diameter of around 4 nm. The current l ₀ carried by charged particles is measured. Then the trap voltage is increased until the measured current I is essentially smaller than l ₀ . This trap voltage is then used in actual mass concentration measurements, as with such trap voltage the sensitivity of the current/mass conversion to changes of median particle diameter is significantly reduced. The trap voltage is preferably increased until the measured current I is 20-95% smaller and more preferably 65-85% smaller than l ₀ .

In one embodiment of the present invention the tarp voltage is set to remove larger nucleation mode particles which are considered to be equal to essentially free ions. This is especially the case within some diesel engine exhaust emission measurements. In general, particle size distributions in diesel engines are classified into three different modes: nucleation, accumulation and coarse modes. In nucleation mode, where the particle diameter is less than 50 nm, CMD of nucleation mode particles being typically less than 30 nm, the particles mainly consist of organic matter, sulfur compounds, and volatile components. Typically monitoring the soot concentration of a diesel engine exhaust requires the removal of the nucleation mode and thus it is considered to be equal to essentially free ions in those cases. Brief description of the figures

In the following the invention will be described in greater detail, in connection with preferred embodiments, with reference to the attached drawings, in which

Figure 1 shows trap penetration, number distribution and mass distribution of a typical single-source aerosol where the particle concentration obeys lognormal distribution; Figure 2 is a schematic view of one embodiment of the invented apparatus; and

Figure 3 shows the relative correlation factor as a function of particle count median diameter with different trap voltages. Detailed description of the invention

Figure 1 shows a typical example of a lognormal particle size distribution. Such particle size distribution with count median diameter (CMD), i.e. the particle diameter where the cumulative number distribution hits 0,5 being around 50 nm and geometric standard deviation (GSD) around 1,7 could well represent particle size distribution from diesel engine exhaust. As seen from the figure, the mass median diameter is almost twice as high as CMD. In Figure 1 the trap voltage is set to a value which corresponds to cut-off diameter of 50 nm. Penetration through the trap is essentially zero for particles smaller than the cut-off diameter and above it the trap penetration smoothly increases.

Figure 2 shows one embodiment of the invented apparatus 1 for particle mass concentration measurement. Apparatus 1 comprises passage 2 with inlet 3 and outlet 3 for guiding sample flow Q comprising particles P, P* with a certain particle size distribution through apparatus 1. The flow Q through passage 2 can be realized in various ways such as by using a pump, by using chimney effect or by using ion wind. In the embodiment of Figure 2, fan 5 drives air into inner passage 2* through filter 6. This air is passed next to the means 7, 8 for ionizing the air. In one embodiment of the present invention, the means 7,8 for ionizing the air are realized by a corona discharge unit 7, powered by a high voltage source 8, which is electrically isolated from mains with an isolation transformer 9. The ionized air forms the motive fluid flow of ejector 10 placed inside passage 2. The ejector generates an underpressure which drives sample flow Q with different size particles P, P* into apparatus 1 vial inlet 3. The ionized air and particles are effectively mixed in the mixing zone 11 and thus particles P,P* are charged 12, 12*. Free ions 11, which as described previously may also be very fine charged particles, are removed by an trapping means 13, which essentially is an electrostatic precipitator removing small charged particles due to their high electrical mobility in an electrical field. The necessary trap voltage is generated by a power source 14, which is controlled by means 15 for controlling the trap voltage. In another embodiment of the present invention the trapping means 13 may be connected to means 15 which control the distance of the trapping electrodes and the electrical field strength is adjusted by adjusting the electrode separation without necessarily adjusting the electrical voltage across the trap electrodes. The electrical current carried by particles escaping passage 2 via outlet 4 is measured using means 16, 17 for measuring electrical current carried by charged particles. Although the preferable way to measure the current is to use means 16 for measuring the escaping current, i.e. the electrical current escaping from apparatus 1 with the particles, other current measurement techniques, such as electrodes collecting at least a fraction of the charged particles may be used as well. The measured current is converted to mass concentration value using suitable means 18, which may be situated in apparatus 1 or the conversion may be carried out elsewhere, e.g. by recording the current values and providing the conversion afterwards.

In apparatus 1 the trapping means 13 for trapping essentially all free ions 11 and charged particles 12 having particle diameter smaller than trap cut-off diameter D _c _o, the cut-off diameter D _c _ ₀ being the particle diameter above which penetration through the trapping means 13 essentially deviates from zero, are connected to means 15 for adjusting the trapping means 13 to adjust the trap cut-off diameter D _c _ ₀ . It is essential to the present invention that the trapping means 13 are adjusted to cut-off diameter D _c _ ₀ , which is significantly higher than the diameter of the essentially free ions, i.e. trapping means 13 remove a significant amount of charged particles from flow Q. In one embodiment of the present invention the trapping means 13 is adjusted to trap cut-off diameter D _c _ ₀ of 10 nm or higher.

Preferably apparatus 1 comprises means 18 for converting the electrical current signal to particle mass concentration value. Trapping means 13 are preferably adjusted to different cut-off diameter D _c . ₀ by adjusting the trap voltage.

It is often preferable to be able to adjust the trap voltage during normal operation of apparatus 1. Thus in one embodiment of the present invention apparatus 1 comprises means 19 for controlling the means 15 for adjusting the trapping means 13 on the basis of the output of the means 16,17 for measuring the electrical current carried by said charged particles 12, 12*. In such embodiment of the present invention the trap voltage V is adjusted during the measurement. This may be done when the process producing particles operates in steady state, e.g. when a combustion engine operates with essentially constant torque. First means 15 for adjusting the trapping means 13 set the trap voltage V to a value, which ensures the removal of essentially all free ions 11, e.g. the trap voltage V is set to value V ₀ which corresponds to particle cut-off diameter D _c _ ₀ of around 4 nm or less. The current l ₀ carried by charged particles 12, 12* is measured using current measurement means 16, 17. Then the means 19 for controlling means 15 for adjusting trapping means 13 increase trap voltage V until the current I measured by means for current measurement 16, 17 is essentially smaller than l ₀ . This trap voltage V is then used in actual mass concentration measurements, as with such trap voltage the sensitivity of the current/mass conversion to changes of median particle diameter is significantly reduced. The trap voltage V is preferably increased until the measured current I is 20-95% smaller and more preferably 65-85% smaller than l ₀ . It should be noted that means 19 for controlling means 15 for adjusting trapping means 13 should be understood as functional components which may be realized in various ways and may e.g. situate in the same control unit. The present invention also includes use of particle mass concentration measurement apparatus 1, where apparatus 1 comprising passage 2 with inlet 3 and outlet 4 for guiding sample flow Q comprising particles P, P* through apparatus 1, means 7,8 for electrically charging particles P,P*, means 16,17 for measuring the electrical current carried by said charged particles 12, 12*, trapping means 13 with cut-off diameter D _c . ₀ for trapping essentially all free ions 11 and potentially also charged particles 12 and means 15 for adjusting the trapping means 13. The use comprises adjusting the trapping means 13 in such a way that the electrical current I carried by charged particles 12 and measured by current measuring means 16, 17 is essentially smaller than the electrical current l ₀ carried by charged particles 12, 12* and measured by means 16, 17 when the trapping means 13 is set to trap essentially free ions 11 only. Trapping essentially free ions preferably means that charged particles with diameter less than 4 nm are trapped.

In other words apparatus 1 is used to remove an essential fraction of charged small particles 12, i.e. charged particles 12 with diameter smaller than the cut-off diameter D _c . ₀ of the trapping means 13. In such case only the electrical current carried by charged larger particles 12* is measured. Surprisingly using apparatus 1 in such a way significantly reduces the sensitivity of mass concentration measurement on the particle median diameter.

The preferred use of apparatus 1 comprises using trap voltage V to control the trap cut-off diameter D _c _

Current I is essentially smaller than current l ₀ . In one embodiment of the use of apparatus 1 the trap voltage is adjusted in such a way that the electrical current signal measured by means 16, 17 for measuring the electrical current carried by said charged particles 12, 12* is preferably 20-95% smaller and more preferably 65-85% smaller than the electrical current signal measured by means 16,17 for measuring the electrical current carried by said charged particles 12, 12* when the trap voltage of the means 13 for trapping essentially all free ions and particles with diameter smaller than cut-off diameter D _c _ ₀ is set to trap essentially free ions only. Another embodiment of using apparatus 1 for particle mass concentration measurement comprises adjusting the trap voltage V in such a way that the cut-off diameter D _c . ₀ is essentially equal to the count median diameter (CMD) of particles (P,P*) in sample flow (Q). Essentially equal may mean e.g. that the cutoff diameter D _c . ₀ is within -50% - +100% of the count median diameter of particles P,P* in sample flow (Q).

In a preferred embodiment of the present invention 2, apparatus 1 comprises means 15 for adjusting the trapping means 13 to adjust the trap cut-off diameter Dc-o to a value which is within 25 nm to 100 nm. Yet another embodiment of using apparatus 1 for particle mass concentration measurement comprises adjusting the trap voltage, by means 15 for adjusting the trap voltage of the means 13 for trapping essentially all free ions and particles P with cut-off diameter D _c . ₀ , to trap voltage V ₀ which ensures the removal of essentially all free ions, with the said trap voltage V ₀ measuring current l ₀ by means 16, 17 for measuring the electrical current carried by said charged particles 12, 12*, and increasing the trap voltage V until the measured current I is essentially smaller than l ₀ . Such embodiment preferably comprises removing all charged particles with diameter less than 4 nm during the removal of essentially all free ions. Within such use the trap voltage V is preferably increased until the measured current I is 20-95% smaller and more preferably 65-85% smaller than l ₀ . Such use of apparatus 1 makes it possible to adjust the trap voltage during the normal use of apparatus 1 to reduce the mass concentration measurement uncertainty due to changing median particle diameter.

Figure 3 shows the effect of the invention on the mass concentration measurement using an electrical particle sensor. If we assume that the CMD of the aerosol under measurement is 50 nm, the normalized conversion factor when converting current measurement result into mass concentration value is one, 1. If the trapping means 13 is set to 4 nm trap cut-off diameter D _c . ₀ , the normalized conversion factor changes nearly linearly as a function of the CMD. Without knowing the CMD changes which may happen within the aerosol under measurement, the use of a correction factor determined for 50 nm CMD will lead to erroneous results when measuring aerosol with same origin but different CMD. However, if a higher trap cut-off diameter is used, e.g. 50 nm D _c _ ₀ , the sensitivity of the normalized conversion factor on the CMD is much smaller. Setting the trap cut-off diameter to such a high value essentially reduces the measured current. If the normalized current l ₀ with 4 nm trap cut-off diameter is 1 when CMD is 50 nm, then the normalized current with 50 nm trap cut-off diameter and 50 nm CMD is 0,25, i.e. considerably less than l ₀ . If the CMD decreases to 30 nm, the normalized current is only 0,09. If CMD increases to 70 nm or 90 nm, the normalized current is 0,39 and 0,49, respectively. Figure 3 shows that when the trap cut-off diameter is set equal to the nominal CMD of the aerosol under measurement, e.g. in the situation of Figure 3 to 50 nm (which is a typical CMD for diesel engine exhaust aerosol), then the CMD variation area where the normalized conversion factor stays essentially constant (e.g. within +/- 10%) is wide and allows accurate mass concentration measurement although the CMD changes. The graphs in Fig. 3 are determined numerically and they are based on the properties of a particle sensor of Pegasor Oy, Finland and properties of combustion based aerosol particles described in Maricq and Wu, The effective density and fractal dimensions of soot particles from premixed flames and motor vehicle exhaust, Aerosol Science 35 (2004), pp. 1251-1274. The density of the primary particles is assumed to be 2 kg/dm3 and the fractal dimension f _d is 2,3.

It is apparent to a person skilled in the art that as technology advanced, the basic idea of the invention can be implemented in various ways. The invention and its embodiments are therefore not restricted to the above examples, but they may vary within the scope of the claims.