Field of the Invention

The invention relates to a method for determining the presence of an impurity substance or impurity substances from a liquid sample. The invention relates to such methods, where low detection limit is preferred or required. The invention relates to measuring concentration of an impurity substance or impurity substances. The invention relates to spectroscopic methods for such purposes. The invention further relates to devices used for performing such methods. The invention further relates to spectroscopic devices.

Background of the Invention

Ecotoxic metals, often referred as heavy metals, are harmful to living organisms. Such materials as arsenic, lead, antimony, zinc and cadmium can pollute underground waters if dumped onto the soil. Industrial process and waste waters commonly include trace amounts of heavy metals. Current monitoring of the industrial waters is based on sampling and subsequent laboratory analysis. While being very sensitive such process has a long delay between successive measurements as well as between the sampling and completion of analytic results. Therefore they are not ideal for active monitoring and cannot, for example, give early warnings of heavy metal leakages that can lead to process disruptions and environmental accidents. In order to obtain online information about metal concentrations in industrial processes and sudden fault situations novel measurement approaches are needed. For such real-time monitoring system of major metal concentrations, the limit of detection (LOD) in the order of 0.05-2 mg/l or 0.05-2 ppm is often required.

Laser-induced breakdown spectroscopy (LIBS) is able to perform a simultaneous analysis of several elements with good selectivity. It involves focusing a high intensity laser pulse or a pulse train on the surface or inside of the sample material. The leading edge of the pulse evaporates and ionizes a minute quantity of the sample. Optical power is converted to the kinetic energy of emerged electrons which leads to a rapid growth of free electron density at the focal point and further absorption of the pulse trailing edge. The vapour heats up and expands and more sample material is ablated. After the laser pulse, ions and electrons recombine and the characteristic spectra of sample material atoms are observed due to radiative relaxation while the originated plasma-like cloud cools down. LIBS has been applied to liquid samples in order to detect metals.

However, numerous random processes disrupt the shot-to-shot repeatability of LIBS of water streams. The shock wave created liquid aerosol particles above the water surface scatter and absorb the emitted light and the sequential laser pulse in the analysis of the surface. The liquid particles and splashes are also likely to contaminate the optical components. When vapour is created in the bulk liquid, quenching and pressure broadening occur inside the bubble produced by the expanding vapour, weakening and broadening the emission lines. In addition, the plasma shielding inhibits the trailing edge of the laser pulse to heat the focal volume thus lowering the vapour temperature. The formed bubbles and cavitation bubble oscillations may also interfere with the collection of emitted light. Summary of the Invention

We present a novel measurement scheme where trace impurities in the sample solution are rapidly pre-concentrated by forming a droplet and drying it to a particle before performing the LIBS. The droplet comprises evaporable liquid, such as water or organic liquid, in addition to the aforementioned impurities. To dry the droplet by evaporation, the droplet is levitated using principles of electrodynamic balance (EDB). When the droplet dries, impurities (or an impurity substance) remain in the formed aerosol particle and their (or its) mass concentration increases several orders of magnitude. The LIBS analysis is performed from a single particle using moderate laser pulse energy. The spectrum shows emission lines of the impurity or impurities. At least a part of the spectrum, such as an emission line, is detected and used for determining the presence, amount, or concentration of the impurity substance or substances. In order to control the measurements process, such as the levitation of the droplet or the particle in a controlled space and/or the triggering time of the LIBS laser pulse, the droplet or the particle is illuminated, whereby the droplet or the particle scatters or reflects some scattered light. To control the process, some of the scattered light is detected, and information related to the droplet or the particle is derived from the detected scattered light. This information is used to control the measurement process; such as the levitating electric field and/or the triggering time.

A corresponding method is disclosed in the independent claim 1 . Preferred embodiments are presented in dependent claims 2 to 57. A corresponding device is disclosed in the independent claim 58. Preferred embodiments are presented in dependent claims 59 to 1 1 1 .

Description of the Drawings Fig. 1 shows, in a top view, an arrangement for performing the measurements,

Fig. 2a shows, in a side view, electrodes used to levitate a droplet or a particle,

Fig. 2b shows, in the cross-sectional side view llb-llb of Fig. 2a, electrodes used to levitate a droplet or a particle,

Fig. 2c shows, in the cross-sectional top view llc-llc of Fig. 2a, two electrodes used to levitate a droplet or a particle,

Fig. 2d shows, in a perspective view, a half of the electrodes used to levitate a droplet or a particle, and the particle,

Fig. 3a shows an example of a first measured spectrum of sample liquid having a first impurity concentration and a second measured spectrum of sample liquid having a second impurity concentration ,

Fig. 3b shows an example of calibration data based on the measured data shown in Fig. 3a,

Fig. 3c shows another example of calibration data based on the measured data shown in Fig. 3a,

Figs. 4a-4b show, in a top view, arrangements for performing the measurements,

Figs. 5a-5b show, in a top view, arrangements for performing the measurements and further performing laser fluorescence measurements, and Figs. 6a-6b show, in a top view, arrangements for performing the measurements and further performing Raman spectroscopy measurements. Detailed Description of the Invention

We disclose a pre-concentration method for laser-induced breakdown spectroscopy (LIBS) analysis of liquid solutions, such as water, that converts a droplet of the sample solution into a single particle by drying the droplet at least to some extent. The term liquid refers to a substance that is in the liquid state in the measurement environment. Drying takes place typically within a few seconds. The particle comprises the trace impurities (or an impurity substance) present in the liquid droplet. The particle is trapped using the electrodynamic balance (EDB) technology. The particle may be trapped to a focal point for the exciting laser of the LIBS. The novel measurement principle has been found to have potential for a compact and sensitive online monitoring method of industrial waters. The method has been proven to be fast, and still to have a low detection limit. The method can be applied to measure impurity concentration of any such liquid sample, from which a generated droplet dries when being levitated, preferably at substantially room temperature. Thus, preferably the liquid sample is in liquid form in standard temperature and pressure (at a temperature 20 °C and in a pressure 101 .325 kPa). Referring to Fig. 1 , the liquid sample 200, possibly comprising the impurity substance, is conveyed in a line 100. A side flow can be taken from the line 100 via a second line 105 to a droplet generator 1 10. By keeping the length of the second line 105 reasonable short, the droplet generator 1 10 receives the sample liquid rapidly from the circulation in the line 100. Preferably, no additives are added to said liquid sample, since this may impair the accuracy of the measurements. However, as will be discussed, a marker material can be added to the liquid. If a marker is used, preferably, the amount of the marker is at most 0.06 w% (i.e. 600 ppm) or at most 0.05 w%. Referring to Fig. 1 , in the method, a droplet 210 is generated from the liquid sample 200, using the droplet generator 1 10. The droplet generator 1 10 can be a single droplet generator. The droplet generator 1 10 may comprise a piezoelectric actuator arranged to pump the droplet 210 out of a nozzle of the droplet generator 1 10. The initial size of the droplet may be e.g. from 20 μηη to 1 mm. It is noted that due to surface tension, the droplet is essentially spherical, whereby the size here refers to the diameter.

The droplet 210 or the liquid sample 200 is electrically charged in such a way that the formed droplet 210 is electrically charged (i.e. its net electric charge is not zero). Preferably, at least the droplet 210 is electrically charged with an electrical charger 120. The electrical charger may be e.g. an electrically conducting plate having a hole or holes or an electrically conductive mesh or grid. The droplet 210 may be arranged to move through a hole of the plate. The plate of the electrical charger 120 is arranged at an electric potential relative to the electric potential of the particle generator 1 10. Thus, the particle 210 becomes electrically charged and, because of the electric field, becomes guided towards a space 400 (see also Fig. 2c). The electric field strength between the plate of the charger 120 and the droplet generator 1 10 may be e.g. at least 20 kV/m, such as from 30 kV/m to 400 kV/m. The electric charger 120 may be configured to expose the droplet to such an electric field. The droplet 210 or the liquid 200 may be charged so that the absolute value of the charge of the droplet 210 is from 10 ⁵ to 10 ⁸ elementary charges (e). One elementary charge (e) is 1 .6*10 ^"19 C. The electric charge of the droplet may be positive or negative. Correspondingly, the electric potential of the plate of the charger 120 relative to the droplet generator 1 10 may be positive or negative.

In the space 400, the droplet 210 is levitated using principles of electro dynamic balance (EDB). Referring to Figs. 2a - 2d, at least part of the space 400, wherein the droplet 210 is levitated, is arranged in between a first electrode 310 and a second electrode 320. In Figs. 2, the electrodes 310 and 320 are arranged in such a way, that the space 400 is limited in the vertical direction by the electrodes 310 and 320. The first electrode 310 is arranged at an alternating upper electric potential Utop+Udc using a first wire 315. The second electrode 320 is arranged at an alternating lower electric potential Ubot using a second wire 325. A control unit 350 is configured to produce the alternating upper electric potential and the alternating lower electric potential. The upper potential is produced by an alternating potential (u _top in Fig. 2d) having a zero mean value and a direct current (DC) potential (Ud _C in Fig. 2d). The lower potential is produced by an alternating potential (Ubot in Fig. 2d) having a zero mean value. The control unit 350 may be a part of a control unit arrangement including a processing unit 199. Even if we here refer to an "upper" potential and a "lower" potential, the electrodes 310 and 320 may be arranged also in such a way, that the space 400 is limited in a horizontal direction by the electrodes 310 and 320. Naturally also other orientations are possible. Preferably, the first electrode is above the second electrode, as indicated in Figs. 2, since then the DC potential can be used to compensate for the gravity. Moreover, it is noted, that a DC potential can be used alternatively (or in addition) to produce a part of the lower electric potential. The sign of the DC potential is selected according to whether it is used in connection with an upper electrode or a lower electrode, and according to the sign of the electric charge of the droplet or the particle; to compensate for the gravity and to control the vertical position of the droplet. Because of the control of the vertical position, the DC component may change in time, and also its sign may change in time. These changes are much slower than the AC component of the electric field. The sign determines the electrode, towards which the droplet or particle is being attracted by the DC component, as known to a skilled person.

Typically in EDB, the frequency and phase of the alternating potentials u _t0 p and Ubot are the same. Moreover, their magnitude may be the same. Thus, in an embodiment the alternating potentials are controlled in such a way that difference between the upper potential (utop+Udc) and the lower potential (u _DO t) equals the DC potential Ud _C at all times during levitation. The DC potential may be constant in time for at least one period of the AC potential.

As the droplet 210 has been electrically charged, an electric field imposes a force on the droplet. The DC component of the electric field formed by the upper and lower potentials is adjusted to balance off the gravitational force affecting the droplet 210. The alternating part of the electric field produces a drag force affecting towards the central point of the volume limited by the first electrode and the second electrode. This happens, because the frequency of the alternating electric field is selected in such a way, that the gas, in which the particle or the droplet is levitated, damps the movements of the particle or the droplet as it oscillates around a central point in the space 400, as known to a skilled person from the principles of EDB. The shape of the projection of the surface of second electrode 320 facing the first electrode 310, when projected on a plane having a surface normal parallel to a line from the centre of the first electrode 310 to the centre of the second electrode 320, has a shape of a loop, such as a circle, as depicted in Fig. 2c. The shape of the projection of the surface of the first electrode 310 facing the second electrode 320, when projected on the aforementioned plane has a shape of a loop, such as a circle. In this way, the first and the second electrodes 310, 320 are hollow, such as cylindrical, as depicted in Figs. 2a-2d. Preferably the aforementioned projections are circular, since this provides for a symmetric electric field and makes the control of the electric field easier.

The levitation of the droplet 210 or the particle 210 depends on the electric charge of the droplet or the particle, and also on the properties of the electric field produced by the combination of the control unit 350 and the electrodes (310, 320, and optionally only 330 or both 330 and 340). As the droplet dries, its mass decreases, which results in a need for controlling the measurement process, at least the electric field. However, if the liquid sample 200 is very dirty and/or the droplet is dried only a little, it may suffice to use only a predefined electric field as function of time, and without a feedback from an optical detector 140, such as a photodetector 140. Controlling the electric field based on the detected scattered radiation has the effect that the droplet 210 can be dried to a greater extent, whereby the detection limit of the method becomes better, because a much greater part of the droplet 210 can be evaporated during levitation. Thus, preferably at least the levitating electric field, i.e. the electric field used to levitate the droplet 210 or the particle 210, is controlled. More preferably, at least the frequency of the levitating electric field is controlled.

In addition or alternatively, the triggering time can be controlled. This has two effects. First, if the droplet 210 is not fully dried, the proper triggering time can be determined by determining the size of the particle. Some sample liquids 200 may be so pure that they cannot be fully dried, since the corresponding droplet 210 would not comprise a reasonable amount of impurity material to form a dried particle 210 of a reasonable size for stable levitation. Second, even if the sample liquid 200 is so dirty that it can be fully dried, whereby the levitation may last for e.g. hours or days, the scattered light can be used to determine a proper triggering time to decrease the measurement time. I.e. the fully dried particle needs not to be levitated for safeguarding sufficient drying, since the sufficiency of drying can be monitored, and time can be saved. When levitating a droplet 210 or a particle 210 that has the properties specified elsewhere, the maximum absolute strength of the electric field used to levitate the droplet 210 or the particle 210 may be e.g. at least 100 kV/m. The maximum absolute strength of the field can be calculated as the ratio of the maximum of the absolute difference between (i,a) the upper electric potential or (i,b) the lower electric potential and (ii) the ground (e.g. the third or fourth electrode 330, 340) to the distance d _e 2 (see Fig. 2c) between the first 310 or the second 320 electrode and the closest ground electrode (330 or 340). Referring to Fig. 2d, the maximum absolute strength may be e.g. |Udc ⁺ uto _P | de2, wherein the potentials Ud _C and u _top are functions of time and the maximum refers to a selection of a point of time. The maximum absolute strength of the electric field may be e.g. at most 3000 kV/m. The control unit arrangement (350, 199) may be configured accordingly.

The direct current (DC) component, Udc, may be selected in such a way that the corresponding DC electric field strength, |Ud _C |/d _e i (see Fig. 2b), is from 6 V/m to 6.5 kV/m. The distance d _e i may be e.g. from 4 mm to 20 mm; such as 8 mm as an example. Thus, the absolute value of the DC component |Ud _C | may be e.g. from 0.05 V to 500 V. The sign of the DC component may be selected with respect to the sign of the electric charge of the droplet in such a way that gravity is compensated. Typically, the droplet is charged by removing electrons therefrom, whereby the droplet becomes positively charged. Thus, the DC component of the potential of the electrode that is above the particle or droplet is negative, at least from time to time, i.e. when the particle is attracted upwards e.g. to compensate for the gravity.

The control unit arrangement (350, 199) may be configured accordingly. The orientation of the direct current component is selected, with respect to the electric charge of the droplet 210, in such a way that the electrostatic force is directed upwards to compensate gravitation. Moreover, as the droplet 210 dries, it loses some mass, whereby the DC component of the field can be reduced while the droplet is levitated. The DC component here refers to time average of the potential Ut _op ⁺ Udc used to generate the electric field, as averaged over at least one period (i.e. period of repetition of the alternating field). Typically, if a low frequency can be used to levitate the particle, then the droplet is (initially) large, whereby long drying time is needed. Therefore, the DC component needs not to be rapidly changed. Thus, in practice the distinction between the AC and DC components of the electric field is evident, even if the DC component may be reduced during levitation.

In the following, it is assumed that the first electrode 310 is arranged above the second electrode 320. An electrodynamic drag force, which will be present when an alternating current (AC) field is used, drives the droplet or particle towards the central point of a line connecting (i) the centre of the such a part of the first electrode 310 that is close to the second electrode 320 and (ii) the centre of such a part of the second electrode 320 that is close to the first electrode 310. This happens, because the alternating electric field produced by the hollow first and second electrodes has such a frequency that the surrounding medium (i.e. gas) decelerates the particle or the droplet because of its drag (i.e. resistance to movements). In Fig. 2b, the droplet 210 is shown in such a central point. Thus, the electric field is an alternating electric field and has an alternating current (AC) component. The AC component results from the alternating upper and lower potentials. For the aforementioned initial size of the droplet 210, the frequency of the electric field is preferably from 10 Hz to 5 kHz. Such frequencies are applicable also for the dried particle, of which size, will be discussed. The control unit arrangement (350, 199) may be configured accordingly. Typically, the droplet 210 or the particle 210 can be levitated in a more stable manner, if the frequency of the electric field is increased during the levitation. This is, because the mass of the electrically charged droplet reduces while drying, whereby it responds faster to the force due to the electric field. The control unit arrangement (350, 199) may be configured accordingly. As evident to a skilled person, the AC field strength also affects the response time of the droplet to the electric field.

As indicated above, the levitation is controlled by the electric field formed by the first and the second electrode, respect to a ground electrode (e.g. 330 or 340). As indicated in Figs. 2, two ground electrodes 330, 340 may be used. The levitation may be better controllable, if a third electrode 330 and a fourth electrode 340 are used and electrically grounded with wires 335, 345, as shown in Figs 2a-2d. The third electrode 330 may be hollow. At least a part of the first electrode 310 may be arranged inside the third electrode 330. The fourth electrode 340 may be hollow. At least a part of the second electrode 320 may be arranged inside the fourth electrode 340. The space 400 may be left in between the third electrode 330 and the fourth electrode 340.

As both the third 330 and the fourth 340 electrodes may be grounded, a single ground electrode, i.e. a third electrode, may suffice. The ground electrode would be electrically grounded with a wire 335. In that case, at least a part of the first electrode 310 and at least a part of the second electrode 320 may be arranged inside the ground electrode. Moreover, the space 400 may be left in the interior of such a ground electrode. As indicated in Figs. 2, a ground electrode (330, 340) radially surrounds at least part of the first or the second electrode. E.g. a common ground electrode may radially surround at least part of the first electrode and at least part of the second electrode. E.g. a third electrode 330 may surround at least part of the first electrode 310 and a fourth electrode 340 may surround at least part of the second electrode 320. The electrodes 330, 340 may surround the part radially. The term "radially" is here interpreted to mean a direction in such a plane, where the intersection of the corresponding first or second electrode and the plane has the form of a loop, e.g. is circular.

While being levitated in the space 400, the droplet 210 dries. When the droplet 210 dries, the impurities remain in the formed particle 210 and thus the mass concentration of the impurity substance (or substances) increases. The particle needs not to be totally dry; it suffices that most of the liquid of the droplet 210 is evaporated, as will be detailed below. Moreover, because of this drying, less optical energy is needed to break the particle 210 to at least some atoms and or ions than would be needed to break the droplet 210 of its initial size.

Droplets 210 are known to dry, when the partial pressure of the corresponding vapour (i.e. the main substance of the liquid sample 200 in the vapour form) surrounding the droplet 210 is less than the saturation vapour pressure of the liquid of the droplet. For water, the saturation vapour pressure at 20 °C is about 2 kPa, i.e. about 0.02 atm. The drying process can be accelerated by lowering the pressure and/or increasing the temperature. Preferably, the measurements can be made at room temperature and in atmospheric pressure. In these conditions, the liquid substance should be in liquid form. Thus, preferably, the boiling point of the liquid is at least 50 °C or at least 60 °C. Thus, preferably, the freezing point of the liquid less than 30 °C and more preferably less than 10 °C. Moreover, to ensure drying, the saturation vapour pressure should preferably be reasonable. Preferably, the saturation vapour pressure of the liquid at 20 °C is at least 0.1 kPa or at least 0.5 kPa. Preferably, the liquid sample 200 comprises at least 95 w% or at least 99 w% water. In an embodiment, the liquid sample 200 comprises at least 95 w% or at least 99 w% alcohol. In an embodiment, the liquid sample 200 comprises at least 95 w% or at least 99 w% other organic constitutes obtainable or obtained by distillation, such as gasoline. The liquid sample 200 may comprise at least 95 w% or at least 99 w% of a mixture consisting of one or more of the aforementioned liquids. The space 400, wherein the droplet 210 (later particle 210) is levitated can be heated to speed up the drying by vaporization. The corresponding device comprises a heater 420 (see Fig. 4). The heater may be arranged to heat the space 400. The heater may be arranged to heat the liquid, e.g. the pipeline 105 or the droplet generator 1 10.

When the droplet 210 is formed, it has an initial mass m ₀ and an initial size (size in the meaning of diameter). As the droplet 210 dries, some of the liquid is evaporated, whereby the mass and the size decreases. Finally, i.e. at the time when the content of the dried particle 210 is analysed by LIBS, the mass of the particle 210 is m _f . In an embodiment, the droplet 210 is dried in such a way that the ratio m _f /m ₀ of the mass of the particle 210, m _f , to the initial mass of the droplet 210, m ₀ , is at most 0.25. Preferably this mass ratio m _f /m ₀ is at most 0.1 , and more preferably at most 0.05 or at most 0.01 . The size of the particle 210, at the time when the content of the dried particle 210 is analysed by LIBS, is preferably from 1 μηη to 20 μηη. Again, due to surface tension, the particle 210 is essentially spherical, and the size refers to diameter.

However, in practice, the size of the particle or droplet can be easily monitored from the detected scattered light, while the mass is less easy to detect. For example, the area of the image of the droplet provides information on the diameter and/or the volume of the droplet or the particle. Such images are commonly circular, whereby the diameter can be directly observed. An embodiment comprises determining an initial volume V ₀ of the droplet using the detected scattered light. Moreover, the volume of the particle before breaking it by LIBS is V _f . In an embodiment, the droplet 210 is dried in such a way that the ratio V _f A/o of the volume of the particle 210, V _f , to the initial volume of the droplet 210, V ₀ , is at most 0.25. Preferably this volume ratio VfA/o is at most 0.1 , and more preferably at most 0.05 or at most 0.01 . Such a ratio can be ensured by using a sufficiently long drying time, or by monitoring the size of the droplet or the particle as it dries, and calculating the corresponding volume. The droplet may be dried (i.e. levitated) e.g. for some seconds, as indicated elsewhere.

Determining the initial volume has the further benefit, that the volume can be taken into account when calculating the concentration. In particular, if the droplet generator generates droplets, of which size varies from one droplet to another.

The droplet 210 or the particle 210 may be levitated in the space 400 e.g. for at least 100 ms, at least 1 s or at least 5 s to dry the particle before triggering the laser pulse. The proper time depends on temperature, pressure, and the substance of the liquid sample 200. The particle 210 is typically sufficiently dry after levitating for a few (1 to 5) seconds. If needed, the droplet may be levitated for days. Preferably, as discussed above, the size of the droplet is monitored, and the triggering time for the LIBS is determined based on the measurements done with the optical detector 140.

Preferably, the droplet 210 is levitated in a reasonable small space 400. In an embodiment, the volume of the space 400 is at most (1 .5 cm) ³ , i.e. 3.4 mil I il itres. In another embodiment, the volume of the space 400 is less than 1 cm ³ , i.e. less than 1 millilitre. The droplet 210 or the particle 210 remains in the space 400 throughout the levitation. Naturally, the droplet 210 is also guided to the space, as depicted in Fig. 1 a. However, this translational movement is not considered levitation, since there the position of the droplet 210 is not controlled with an alternating electric field. The levitating and alternating electric field spreads to said space 400. Moreover, preferably, for some time, say 10 ms, before triggering the laser pulse 152, the position of the particle 210 is substantially constant. Thus, in an embodiment, the particle is levitated in the space 400 in such a way, that a maximum linear dimension of such a part of the space 400, where the center of the particle 210 is located for the last 10 ms before the triggering time, is at most 40 μηη or at most 20 μηη. In addition or alternatively, the volume of that part of the space may be e.g. at most 8000 μηη ³ or at most 4500 μηη ³ . The particle 210 may oscillate within this small part of the space 400, whereby its position is essentially constant. This helps to focus the laser pulse correctly. As will become clear, the aforementioned part of the space 400 may be arranged at a focal point of optics focusing the LIBS laser pulse 152.

To control the measurement process, especially (i) the position of the droplet 210 or the particle 210 and/or (ii) the triggering time of the laser 150 source, the droplet 210 and/or the particle 210 is optically monitored during drying the droplet to the particle. The measuring process can be controlled using information gathered by said monitoring. To this end, the droplet 210 or the particle 210, which is levitated in the space 400, is illuminated with light 132 using a light source 130 (see Fig. 1 ). The light source 130 may be a laser light source. The light source 130 may be configured to emit light 132 at least at the wavelength range from 400 nm to 900 nm. Thus, the droplet 210 or the particle 210 scatters or reflects some of the illuminating light 132 as scattered light 134. The scattered light 134 is detected using an optical detector 140. Figure 1 shows also means 142 for sending a signal from the optical detector 140 to the control unit 350. The control unit 350 is configured to control the measurement process based on such a signal 142. The measuring process can be controlled by controlling at least one, preferably both, of: (i) the electric field and (ii) the triggering time, i.e. the time when the LIBS laser pulse is triggered. Preferably, at least the electric field is controlled, as discussed above. With reference to Fig. 2d, the electric field can be controlled by controlling at least one of (i,a) the frequency of the electric field, i.e. the frequency of the alternating part u _top and Ubot, (i,b) the DC component of the electric field, i.e. the voltage Ud _C , and (i,c) the amplitude of the alternating electric field, i.e. the amplitude of u _top or Ubot- Preferably, at least the frequency of the electric field is controlled. As evident, this controlling can be done automatically by the control unit 350. Thus, the upper and lower potentials of the first and the second electrode 310, 320 can be controlled relative to each other. Moreover, if a ground electrode 330 is used, the upper and lower potentials (dc part, frequency, and/or amplitude) of the first 310 and the second 320 electrodes can be individually controlled relative to electric ground. Typically u _top and Ubot have the same frequency and phase. When controlling the measurement process, information related to the droplet 210 or the particle 210 is derived from the detected scattered light 134. Such information may include one or more of: the size of the droplet 210 or the particle 210 and the position of the droplet 210 or the particle 210. Moreover, these can be detected as function of time (i.e. in at least two instances of time), whereby the degree of drying can be detected by comparing the present size with the initial size (i.e. the change in size). In addition, velocity can be determined as a time derivative of the position. The measuring process can be controlled by using such derived information related to the droplet or the particle. In particular,

- the electric field can be controlled using the position and/or the size of the droplet or the particle,

- the electric field can be controlled using the positions and/or the sizes at two instances of time, the droplet or the particle, e.g. by calculating a velocity and/or a rate of evaporation,

- the triggering time can be controlled using the position and/or the size, of the droplet or the particle and/or

- the triggering time be controlled using the positions and/or the sizes at two instances of time, the droplet or the particle, e.g. by calculating a velocity and/or a rate of evaporation.

In the method, a laser pulse 152 is generated using a laser pulse source 150. The laser pulse 152 is generated at a triggering time, in response to a triggering signal 355 received from the control unit 350. The laser pulse 152 is capable of breaking down the particle 210, which has been obtained from the droplet 210 by drying at least to some extent, and is thus also capable of producing some atoms and/or ions into the space 400. As for the capability, the laser pulse 152 may be focused to a focal point 410 to increase its intensity. The laser pulse 152 is guided to the space 400, wherein the particle 210 is being levitated; and on to the levitating particle 210. Thus, the laser pulse 152 hits the particle 210 and breaks it down. A leading edge of the laser pulse 152 atomizes and/or ionizes at least some material of the particle 210, and the trailing edge of the pulse 152 heats the atoms and/or ions atomized by the leading edge, by absorption of light into the atoms, ions, and/or free electrons. These atoms, ions, and/or free electrons may recombine to atoms and/or ions. Because the temperature is high, the atoms and/or ions may be in an excited state. In this way, the some atoms and/or ions become thermally excited. Because the atoms are thermally excited, the method is reasonably insensitive to the wavelength of the laser pulse, unless indicated otherwise in connection with some embodiments.

When these atoms/ions de-excite, they emit electromagnetic radiation 154 including or consisting of light. The spectrum of this radiation 154 is specific to the composition of the particle 210. At least a part of the spectrum of the radiation 154 specific to the composition of the particle 210 is detected using a spectrometer arrangement 160. The spectrometer arrangement 160 is configured to detect at least a part of the spectrum of radiation 154 emitted by the atoms and/or ions of the broken particle, which particle 210 before the breaking was arranged in the space 400, optionally in a focal point 410. A spectrum may comprise e.g. multiple emission peaks at specific wavelengths. A part of such a spectrum may comprise e.g. only one peak. It may be sufficient to monitor only a part of the spectrum (i.e. a detected part of the spectrum), such as a predefined peak of the spectrum. As is conventional, the term spectrum refers to a continuum of wavelengths in connection with an intensity for all wavelengths of the continuum of wavelengths. In what follows, the term "detected spectrum" is used to clarify this issue whenever that seems appropriate.

Moreover, in this description, the "detected spectrum" includes an intensity or a spectral intensity form at least one wavelength; such as an intensity of a peak of an emission spectrum. The terms intensity and spectral intensity refer to the radiation on the optical detector 140. These spectral properties (i.e. those that are here called intensity and spectral intensity) are often referred to as irradiance and spectral irradiance, respectively. However, at high temperatures, peaks of emission spectra are known to widen. Thus, instead of spectral intensity Ι(λ) at a given wavelength or at a range of given wavelengths, a total intensity of a broader band, ί\{λ)όλ, can be used. Herein the integral is taken over the aforementioned band. The band may comprise all the relevant wavelengths for a (broad) emission peak. Either of these intensities, the spectral intensity Ι(λ) or the total intensity ί\(λ)όλ, is usable for deternnining the presence or concentration or amount of the impurity substance. Moreover, the intensity (in either meaning) can be given in any units, such as in proportion to a noise ratio, in photon counts (per time and per wavelength for the spectral intensity or per time for the total intensity) of an optical detector, in W/m ² /m (spectral intensity), or in W/m ² (total intensity). In practice, the spectrometer arrangement measures intensity over a certain wavelength band, from which the spectral intensity can be calculated.

When determining the presence of the impurity, the value of the intensity or spectral intensity is not important, as long it exceeds a limit. However, when interested in the amount or concentration, also the intensity or spectral intensity counts. Therefore, in an embodiment an intensity signal is formed using the detected part of the spectrum of radiation specific to the composition of the particle. The intensity signal is indicative of at least one of - the total intensity Jl( )d of the detected part of the spectrum of radiation specific to the composition of the particle, and

- at least a spectral intensity Ι(λ) of the detected part of the spectrum of radiation specific to the composition of the particle at a wavelength λ comprised by the detected part of the spectrum.

Such an intensity signal may be e.g. one of the following:

- the spectral intensity Ι(λ) as given in W/m ² /m,

- the total intensity for a selected band ί\{λ)όλ, as given in W/m ² ,

- an output signal of the spectrometer arrangement, e.g. in units of V or photon counts (see Fig. 3a),

- an output signal of the spectrometer arrangement relative to a noise level thereof (see Fig. 3b),

- an output signal of the spectrometer arrangement corresponding to an emission line of the impurity, relative an output signal of the spectrometer arrangement corresponding to an emission line of a marker (see Fig. 3c).

It has been found that for the aforementioned size of particle 210, the energy of the laser pulse 152 is preferably from 0.1 mJ to 100 mJ. The duration of such a pulse 152 may be from 10 fs to 100 ns. The wavelength of the laser pulse 152 is preferably from 200 nm to 2000 nm. It may be e.g. 355 nm, 532 nm, or 1064 nm. The light pulse source 150 may configured accordingly. Light having shorter wavelength is significantly absorbed by the components of air. Thus, if a shorted wavelength is used, the environment is preferably also changed to an inert atmosphere comprising nitrogen, helium, and/or neon; such as to an inert atmosphere consisting of nitrogen, helium, and/or neon. As for longer wavelengths, such a laser pulse does not easily break down the particle 210, whereby much larger laser pulse energies might be needed. In an embodiment, the droplet, while being levitated, is surrounded by an inert atmosphere, such as an atmosphere comprising at least 90 vol%, preferably at least 95 vol%, nitrogen (N ₂ ), helium (He), neon (Ne), argon (Ar), or a mixture thereof. Moreover, the water comprised by air may absorb some emitted light, provided that the wavelength correspond to absorption of water. This issue can be overcome by using dried air as the atmosphere. Dried air may have a relative humidity of e.g. at most 10 %RH or at most 5 %RH. The dried air may be substantially dry, e.g. comprising at most 100 ppm water. Alternatively or in addition, the droplet and/or the particle may be levitated in a substantially vacuum environment, such as in an environment having a pressure of at most 10 kPa or at most 1 kPa.

Preferably, the laser pulse 152 is focused to a focal point 410 using focusing optics 500, such as a lens 155 or a reflector configured to focus the laser pulse 152 to the focal point 410. The lens 155 may be configured to focus the light pulse 152 in such a way that the diameter of the laser pulse at the focal point 410 is from 0.5 μηη to 100 μηη. The diameter refers to full width at half maximum (FWHM); i.e. the diameter of the area, wherein the intensity is at least half of the maximum intensity. Moreover, the diameter refers to the diameter of the cross-section of the pulse 152 on a plane having surface normal parallel to the direction of propagation of the light pulse 152. The focal point 410 is located in the space 400, and the particle 210 is preferably levitated in said focal point 410.

As for the control of the measurements, an embodiment comprises (i) determining that the droplet 210 has dried to the particle 210 and (ii) determining that the particle 210 is at the focal point 410 at a predicted triggering time. Moreover, the laser pulse 152 is generated at the predicted triggering time. In an embodiment, at a first time t _pre d, one determines, using the information related to the droplet or the particle, that the droplet has dried to the particle, and determines a predicted triggering time t _t ng, _p , at which triggering time the particle is predicted to be at the focal point 410. Thereafter, the laser pulse 152 is generated at the predicted triggering time, i.e. the triggering time is set equal to the predicted triggering time. Preferably the time difference between the predicted triggering time ttrig.p and the first time t _pre d, i.e. ttrig,p-t _P red is at most 500 ms or at most 200 ms. This improves the accuracy of the prediction for the triggering time. When a control unit 350 and a laser pulse source 150 are used, they may inherently have some delay. Such a delay may be taken into account. Thus, in an embodiment, the control unit 350 is configured to send the triggering signal 355 to the laser pulse source 150 at such a time, that the laser pulse 152 is generated at the predicted triggering time t _t ng, _p . The control unit may be configured to send, at a second time t _tr ig,s, the triggering signal 355 to the laser pulse source 150. Moreover, the time difference between the second time t _tr ig,s and the first time t _prec i, i.e. ttrig, _s -t _P red may at most 500 ms or at most 200 ms.

After detecting the spectrum of radiation 154 specific to the composition of the particle, the presence of an impurity substance in the particle 210 is detected using the detected spectrum of radiation 154 specific to the composition of the particle 210. The intensity signal, as discussed above, can be used. As an example, Fig. 3a shows a detected spectrum of radiation 154, as detected from a water droplet comprising zinc. The detected spectrum shows clear peaks, approximately at the wavelengths 468 nm, 472 nm, and 481 nm. It is noted that the LIBS emission spectrum of zinc further comprises peaks in the range 250 nm to 330 nm and at 636 nm. However, these peaks are not detected for Fig. 3a. By using (i) the number of observed peaks, and (ii) the relative intensities of the corresponding peaks, the constituent or constituents of the impurity substance (or impurity substances) of the particle 210 can be determined as known to a person skilled in the art of spectroscopy. It is also possible to monitor only a narrow wavelength band of the spectrum and to measure the intensity signal corresponding to only one emission peak of an interesting impurity substance, such as an impurity atom or ion. When an intensity increase at a wavelength (or several wavelengths) is observed, this implicitly means that the corresponding impurity is present. The intensity increase should exceed also the signal to noise ratio of the spectrometer arrangement 160. In an embodiment, a processing unit 199 is configured to receive a signal 165, such as the intensity signal, from the spectrometer arrangement and by using this signal in the aforementioned way, to determine the presence of a impurity substance in the liquid sample using the detected spectrum of radiation 154 emitted by the particle 210 arranged in the space 400. The impurity substance may be an element (i.e. a chemical element). The impurity substance may be a metal. The impurity substance may be a heavy metal.

The spectrometer arrangement 160 may comprise e.g. a filter, configured to pass only a selected wavelength band and a photodetector configured to determine the total intensity of light at that wavelength band. The band may correspond to a peak of the LIBS emission spectrum. In the alternative, the spectrometer arrangement 160 may be configured to measure the intensity of the light emitted from the particle as the results of LIBS laser pulse as function of wavelength, i.e. for multiple wavelengths. Moreover, in particular, when more than one peak is detected, the spectrometer arrangement 160 may comprise e.g. a beam splitter, another filter and another photodetector configured to detect a part of the spectrum of the emitted light 154.

As known to a skilled person, by observing at least two different emission peaks of the spectrum of radiation 154, the presence of also at least another impurity substance in the particle 210 can be determined. The processing unit 199 can be configured accordingly. A warning signal can be generated, e.g. by the processing unit 199, when the presence of an impurity substance has been detected. The warning signal may be indicative of what is the impurity substance of which presence has been observed.

The detected spectrum of radiation 154 can also be used to determine the amount of the impurity substance or the amounts of the impurity substances in the particle 210. The magnitude of the intensity signal corresponding to a specific radiation peak (or specific peaks) of an impurity substance is indicative of the amount that impurity substance. The magnitude of the intensity signal depends also on the energy of the laser pulse 152 in a known way. Thus, from the magnitude of the intensity signal, the amount of the impurity substance can be determined. Moreover, the magnitudes of the intensity signals corresponding to specific radiation peaks of a second an impurity substance is indicative of the amount the second impurity substance. Thus, by observing the magnitudes of the intensity signals specific to two different impurity substances, the amounts of the impurity substances in the particle 210 can be determined.

Moreover, if the initial mass or volume of the droplet 210, as generated by the droplet generator 1 10, is known, the concentration of the impurity substance or the concentrations of the impurity substances can be determined. Even if the mass or volume is not known, the concentration(s) of the impurity substance(s) can be determined by using calibration measurements.

A warning signal can be generated based on the amount or concentration of at least one impurity substance. The determined amount or concentration can be compared to a warning limit, and when the warning limit is exceeded, a warning signal can be generated. In particular, if the presence of at least to impurity substances is monitored, the warning signal can also be indicative of the of the impurity substance, of which amount or the concentration has exceeded the limit. Thus the signal may be indicative which of the monitored impurity substances is such that its amount or concentration exceeds the warning limit. The processing unit 199 can be configured accordingly.

A control unit arrangement may comprise the control unit 350 and the processing unit 199. The control unit arrangement can be integrated in a single entity, such as an integral part, as depicted in Fig. 4. Moreover, also other parts of the arrangement of Fig. 1 can be integrated to from a device for measuring the presence of an impurity substance in a liquid sample.

As for determining (a) the amount of at least one impurity substance in the droplet 210 or particle 210 or (b) the concentration at least one impurity substance in the liquid sample 200, there are at least three possibilities:

(A) adding some marker to the liquid sample 200 in such a way the concentration of the marker in the liquid sample is known,

(B) calibrating the measurements by using (i) a first liquid sample 200 having a first known concentration of a known impurity substance and

(ii) a second liquid sample 200 having a second known concentration of the known impurity substance, or (C) using theoretical calculations for determining the amount from the measured intensity, and, optionally, the concentration from the amount.

In addition, two or all of these possibilities can be used simultaneously to improve the measurement accuracy.

As for (A), when the concentration of the marker in the liquid sample is known, the spectrum of radiation 154 comprises peaks specific to the marker and to the impurity substance(s). This situation is shown in Fig. 3a, where the emission spectra of zinc (Zn) is shown for two different concentrations (50 ppm depicted in grey and 20 ppm in black). In addition, 600 ppm (i.e. 0.6 g/l) of salt (sodium chloride, NaCI) was added as a marker, whereby the sample comprised about 240 ppm Na ⁺ ions. The emission peak of sodium at 475 nm is schematically shown by a broken line. The concentration of zinc is determined from the intensity signal. The intensity signal in this case refers to the height of the strongest emission peak in the spectral intensity Ι(λ). Provided that the peak broadens, e.g. because of excessive temperature, the total intensity 1ΐ(λ)οΙλ could be used as the signal, as discussed above. In the alternative, one can consider the intensity signal to be the ratio of the height of the emission peak at 481 nm (corresponding to Zn) to the height of the emission peak at 475 nm (corresponding to 240 ppm Na ⁺ ).

In principle, the marker can be seen as being another impurity substance, with the exception that its concentration is known. For measuring two impurity substances, the reader is referred to what has been said above. By monitoring (i) the intensity of a peak specific to an impurity substance and (ii) the intensity of a peak specific to the marker, the concentration of the impurity substance can be determined. The ratio of these intensities (or corresponding signal strengths) is indicative of the ratio of the concentrations of the impurity substance and the marker. As the concentration of the marker is known, the concentration of the impurity substance can be determined. Moreover, when the initial size of the droplet 210 is known, the concentration is indicative of the amount of the impurity substance (and/or marker) in the particle 210. The, the amount can also be calculated, if needed. The marker may consist of only one element. The marker may comprise at least two elements. Even in the latter case it would be possible to monitor an emission peak related to only one elements of the marker (see above). The marker may comprise a substance that dissolves into the liquid. Water typically comprises an unknown amount of sodium (Na ⁺ ) and chlorine (CI ^" ) ions an unspecified amount. Thus, preferable sodium (Na), chlorine (CI2), or sodium chloride (NaCI) is not used as a marker, even if shown as such above. Preferable markers include transition metals.

Referring to Fig. 3c, the measurements may be calibrated by measuring the ratio of the signal from the sample (such as zinc at 481 nm) to the signal of the marker (such as 240 ppm Na ⁺ at 475 nm ), and forming calibration data indicative of how this ratio depends on the concentration of the sample, such as zinc. Fig. 3c shows the ratio (i.e. "Zn/Na intensity ratio") as function of zinc concentration, when using the Na ⁺ marker as indicated above. The circles indicate the measured results, and the dotted line is a linear fit to the measured data. Thus, when an intensity ratio for zinc and sodium is measured, the zinc concentration can be determined by using such calibration data.

Thus, this calibration data (later called first calibration data) is indicative of a relation between (i,a) the amount of a specific impurity substance in a droplet 210 or particle 210 or (i,b) the concentration of a specific impurity substance in the liquid sample 200 and (ii) the ratio of the spectral intensities of radiation 154 specific to an impurity substance and a marker substance; from a liquid sample 200 comprising the impurity substance and a known amount of the marker. Alternatively to the ratio of the intensities, a ratio of the magnitudes of the intensity signals may be used. The sensitivity of the spectrometer arrangement may depend on wavelength. The amount or concentration can be determined using spectral intensity of the radiation specific to the composition of the particle (i.e. the intensity of at least one emission peak) and the calibration data.

Referring to Fig. 1 , the arrangement (or device) may comprise means (170, 175) for feeding marker to the liquid 200. As depicted in Fig. 1 , the means may comprise a pump 170 and an outlet 175, such as a line, for feeding the marker to the liquid 200. The outlet 175 may be e.g. connected to the line 105 configured to convey the liquid 200 to the droplet generator 1 10. As for (B), the procedure can first be calibrated alternatively without a marker. While calibrating, at a first time, a first liquid sample 200 having a first known concentration of a known impurity substance is measured as discussed above. This results in first characteristic intensity signal, for at least one characteristic emission peak of the impurity substance. The first time may be e.g. the time measuring the sample having 20 ppm zinc (see Fig. 3a). While calibrating, at a second time, a second liquid sample 200 having a second known concentration (different from the first known concentration) of the impurity substance is measured as discussed above. This results in second characteristic intensity signal, for at least one characteristic emission peak that was measured also at the first time. The second time may be e.g. the time measuring the sample having 50 ppm zinc (see Fig. 3a). From this, one may deduce calibration data (later: second calibration data), which is indicative of a relation between (i) the concentration of the impurity substance and (ii) the intensity signal related to radiation specific to a liquid sample comprising the impurity substance. The second calibration data may be e.g. a linear dependence between the concentration and the intensity signal for an emission peak. Such calibration data is shown in Fig. 3b. The circles indicate measured results, and the dotted line is a linear fit to the measured data. In Fig. 3b, the intensity signal is given in proportion to the noise, i.e. as a signal-to-noise ratio (SNR). Noise, on the other hand is related to the deviation of the measured intensity, when the sample is free from the substance. The calibration measurements results for 50 ppm and 20 ppm are indicated with grey and black arrows, respectively. From such data, also the limit of detection (LOD) can be determined. The limit of detection is the largest undetectable concentration. A concentration is detectable, when the signal from such a sample exceeds noise; and corresponding is undetectable, when the magnitude of the signal is at most the noise. Thus, the limit of detection can be calculated as the concentration, when the SNR of the signal equals one.

Fig. 3b relates to the 481 nm emission peak of zinc. Preferably, the strongest emission peak is used, as this gives better results in terms of LOD (i.e. the LOD becomes smaller for such a peak). In principle, different linear dependencies can be used for different spectral peaks. Moreover, many different known concentrations can be used for calibration measurements and/or a more complex functional form can be fitted to make a functional relation between an intensity and a concentration. During calibrations, the base liquid for the different concentrations is preferably the same as the liquid, from which the impurity substance(s) is/are to be measured. When measuring an unknown liquid sample, such calibration data (i.e. second calibration data) can be received and used. By determining the magnitude of the intensity signal of the detected spectrum of radiation 154 at a wavelength, the calibration data can be used to calculate, using the functional form (e.g. linear form), the concentration from the magnitude of the intensity signal. Several intensity signals corresponding to different peaks can be used to improve the accuracy.

Alternative to concentration, the calibration may be done based on the amount instead of concentration, e.g. when the initial size of the particle is known.

Thus, the second calibration data is indicative of a relation between (i,a) the amount of a specific impurity substance in a droplet 210 or particle 210 or (i,b) the concentration of a specific impurity substance in the liquid sample 200 and (ii) the magnitude of an intensity signal specific to a liquid sample 200 comprising the impurity substance. The amount or concentration can be determined using the measured intensity signal and the calibration data.

As for (C), the amount of the impurity substance can be calculated using theoretical considerations. To this end, the temperature of the atoms and/or ions can be solved by studying the intensities of two emission peaks. It is known that the intensity of a specific emission peak is indicative of the amount of the impurity substance in the particle 210, and also depends on the temperature. Moreover, the sensitivity of the spectrometer arrangement 160 plays a role in the calculations, and this can be taken into account by measuring an intensity signal at at least two emission peaks. When the amount of the impurity is known, the concentration can be calculated when the initial size of the droplet 210 is known. The initial size may, in some cases, be calculated with reasonable degree of accuracy from the detected scattered light, as described elsewhere. For example, the diameter is indicative of the volume, as the droplet is spherical. As for the calculations, all the elements (such as the pulse energy) affecting the intensity, except for the amount of the impurity substance, are preferably constant from one measured particle to another. Moreover, they may be known. Also preferably, in case this approach is used, a marker is also used. In the calculations, the energy of the laser pulse 152 is preferably also known, because this may further increase the accuracy of the calculation. To measure the amount of optical energy that is used to break down the particle and excite the resulting atoms and/or ions, a second optical detector 145, such as a photodetector, can be used, as depicted in Figs. 4a and 4b. A filter 146 may be used to pass only the remaining part of the laser pulse 152 to the second optical detector 145, and, conversely, to prevent other light, such as the illuminating light 132 from entering onto the second optical detector 145 (see Fig. 4a). However, such a second optical detector 145 can be used that not filter is needed.

The measurement process can be performed subsequently to several individual droplets. As a result, the concentration can be calculated as a statistical measure (e.g. average or median) of the results from multiple droplets. This in general improves accuracy and lower the detection limit. Moreover, the measurements become easier, provided that the droplet generator is arranged to produce droplets having the same size, at least with reasonable accuracy. Preferably, the standard deviation of the initial droplet diameter is at most 10 % of the average initial diameter. In a corresponding embodiment, a second droplet is generated from said liquid sample. A second particle is form by drying the second droplet, and the particle is analysed by EDB and LIBS, as discussed above. Moreover, the measurement process is controlled in a way discussed above. As a result of the laser pulse, the constituents of the broken second particle emit a second spectrum of radiation specific to the composition of the second particle, which second spectrum of radiation is detected. Thus, (i) the amount of at least one impurity substance in the second particle or (ii) the concentration of at least one impurity substance in the liquid sample can be determined using both (a) the detected spectrum of radiation specific to the composition of the particle (as discussed above for the particle) and (b) the second detected spectrum of radiation specific to the composition of the second particle. A corresponding device or arrangement can be configured correspondingly. In particular the particle generator 1 10 arrangement can be configured to generate a sequence of droplets. The control unit 350 may be configured to trigger, at a first triggering time, a first laser pulse to break the first particle, and to trigger, at a second triggering time, a second laser pulse to break the second particle. Preferably, only one particle at a time is levitated in said space 400.

As is evident, preferably multiple single particles are generated and analysed subsequently to further increase the measurement accuracy. A corresponding device or arrangement can be configured correspondingly.

As for the optics of the device or arrangement for the measurements and with reference to Figs. 4a and 4b, the device may comprise first collecting optics 510, such as a lens, configured to focus some of the radiation 154 emitted by the atoms and/or ions of the broken particle to the spectrometer arrangement 160. In addition, the device may comprise second collecting optics 515 configured to collect some of the radiation 154 emitted by the atoms and/or ions of the broken particle to the spectrometer arrangement 160 via the first collecting optics 510. In the alternative, only a reflector (similar to the second collecting optics 515 of Fig. 4) can be used as the first collecting optics. A major portion of the emitted radiation can be collected in such a configuration, where at least a part of the space 400 is left in between the second collecting optics 515 and the first collecting optics 510.

The device preferably comprises collection optics 520, such as a lens 520, configured to collect some of the scattered radiation 134 to the optical detector 140. The device may comprise an interference filter 525 arranged in between the space 400 and the optical detector 140. The purpose of the filter 525 is to protect the optical detector 140 from e.g. daylight and/or the light emitted as a result of the LIBS pulse. The interference filter may be arranged in between the center of the space 400 and the optical detector 140. The interference filter may be arranged in between the optical detector 140 and the collection optics 520 configured to collect some of the scattered radiation 134 to the optical detector 140. As depicted in Fig. 4a, the device may comprise an optical element 530 configured to guide the laser pulse 152 and the illuminating light 132 onto a same optical path. However, as depicted in Fig. 4b, such an element is not necessary. As indicated in Fig. 4b, the device may comprise an intensifier, such as a photomultiplier tube 540, configured to intensify the scattered light 134; whereby the optical detector 140 is configured to detect the intensified scattered light. The intensifier 540 may be an integral part of the optical detector 140, or a separate photomultiplier tube 540 can be used.

The optical detector 140 may comprise a camera, such as a complementary metal-oxide semiconductor (CMOS) camera or a charge-coupled device (CCD) camera. The optical detector 140 may comprise photodiode, such as a segmented photodiode.

In addition to LIBS analysis of the particle, other non-destructive spectroscopic analysis can be performed to the droplet or the particle before the LIBS analysis. Non-destructive refers to such methods, where at least a major portion of the droplet or particle remains in liquid or solid form. In such non-destructive analysis, material of the droplet or the particle is optically excited. As indicated above, in LIBS, the material is thermally excited. Such methods include laser fluorescence and Raman scattering.

These analysis methods can be performed to the same droplet or particle, to which the LIBS analysis is applied. In the alternative, these analysis method can be performed to another particle formed from the same liquid sample 200 using the droplet generator 1 10. Thus, the droplet 210, the particle 210, the other droplet or the other particle may be levitated in the space 400 as discussed above, and analysed by using spectroscopic techniques, such as laser fluorescence or Raman spectroscopy. Because of the difference in the aforementioned methods, the non-destructive methods typically are sensitive to the wavelength of the exciting light. However, as LIBS excites material thermally, the wavelength of the LIBS laser is less important.

With reference to Figs. 5a to 6b, the droplet 210, the particle 210, another droplet or another particle is levitated in the space 400, in an electric field, and for a time. Laser light (182, 192) is generated and guided onto the droplet 210, the particle 210, the other droplet or the other particle. Because of the light 182, 192, material of the droplet or the particle is optically excited. Thus, the droplet 210, the particle 210, the other droplet or the other particle emit some light (184, 194). A spectrum of the emitted radiation can be detected, and the presence of other impurities in the liquid sample can be determined using information derived from the spectrum of the emitted radiation (182, 194). Spectroscopic techniques can be used for the purpose.

Laser fluorescence may be used to determine the presence of other impurities, such as impurities of microbiological origin like bacterium, fungi, or spore, in the liquid sample. The concentration of these may be so low, or the size of these so large, that typically at most one bacteria, fungus, or spore is present in the droplet. Presence can be measured, if at least one bacteria, fungus, or spore is detected. Concentration can be measured by counting the proportional number of droplets wherein such impurities of microbiological origin are present to the total number of droplets measured.

With reference to Figs. 5a and 5b, when using laser fluorescence, a low- power laser pulse 182 is generated. As depicted in Fig. 5a, a low-power laser pulse source 180 can be used for the purpose. As an alternative to a separate low-power laser pulse source 180, the low power laser pulse can be generated from the laser pulse 152. To reduce the power to acceptable level, a skilled person has multiple options. For example, as depicted in Fig. 5b, a reflector 551 , or a combination of reflectors 551 , 552, 553 can be used to form the low-power laser pulse. Provided that the reflectivity is low enough, the power of the pulse is also low enough. E.g. the reflection from a window may provide for suitably low reflectivity. A shutter 550 can be used to prevent the laser pulse 152 from breaking down the droplet. The shutter 550 and the reflector 551 can be integrated to a single component. As an alternative to the reflections, the shutter 550 may comprise an aperture. Thus, this aperture may pass a suitable amount of light for the fluorescence measurements. When the aperture is opened or the shutter removed, the LIBS pulse 152 may propagate to the space 400. The low-power laser pulse 182 may be guided to the space 400 from any direction; compare to the illuminating light 132 of Fig. 4b.

The low-power laser pulse 182 is guided to the droplet 210, the particle 210, the other droplet, or the other particle, which is being levitated in the space 400, whereby the constituents of the droplet 210, the particle 210, the other droplet, or the other particle are optically excited by the low-power laser pulse 182. As indicated above, in contrast to LIBS, the power of the low-power laser pulse 182 is selected such that the low-power laser pulse 182 does not break down the droplet 210, the particle 210, the other droplet, or the other particle, even if guided on to it (i.e. even if guided on to the droplet 210, the particle 210, the other droplet, or the other particle). Moreover, provided that the low-power laser pulse 182 is made from the laser pulse 152, the wavelength of the pulse 152 should be selected in accordance with fluorescence measurements. When the constituents of the droplet 210, the particle 210, the other droplet, or the other particle to de-excite, they emit radiation 184 by fluorescence. This radiation 184 is be detected, e.g. by the spectrometer arrangement 160, or in principle using another spectrometer arrangement. In addition, the presence of other impurities, such as impurities of microbiological origin like bacteria, fungi, or spore, in the liquid sample 200 is determined using information derived from the radiation 184 emitted by fluorescence. In a corresponding device, the spectrometer arrangement 160 or another spectrometer arrangement is configured to detect radiation 184 emitted by said fluorescence. Moreover, the control unit arrangement (350, 199) is configured to determine the presence of the other impurities, such as impurities of microbiological origin like bacteria or spore, in the liquid sample using information derived from the radiation 184 emitted by fluorescence.

Concentration of the other impurity may be determined as indicated above. The control unit arrangement (350, 199) may be configured accordingly.

Referring to Figs. 6a and 6b, in a similar manner, Raman spectroscopy may be performed to the droplet 210, the particle 210, another droplet, or another particle. In Raman spectroscopy, laser light 192 (continuous or pulsed) is used. The laser light 192 may be generated with a laser light source 190 (see Fig. 6a), or in a similar manner as discussed above for the laser fluorescence (see Fig. 6b), because pulsed light can be used also for Raman spectroscopy. The power of the laser light 192 is selected such that the laser light 192 does not break the droplet 210, the particle 210, the other droplet, or the other particle, even when guided on to it. The laser light 192 is guided to the droplet 210, the particle 210, the other droplet, or the other particle, which is being levitated in the space 400. Thus, the constituents of the droplet 210, the particle 210, the other droplet, or the other particle inelastically scatter (i.e. Raman scatter) some of the laser light to Raman radiation 194. A spectrum of the Raman radiation is detected, and the presence of other impurities in the liquid sample 200 is determined using information derived from the Raman radiation. Concentration of the other impurities, or the amount thereof is a droplet or a particle, can be calculated using similar techniques as discussed above for LIBS. Moreover, provided that the lase light 192 is made from the laser pulse 152, the wavelength of the pulse 152 should be selected in accordance with Raman measurements.