WO2003001190A1

WO2003001190A1 - X-ray fluorescence analyser, and a method for using an x-ray fluorescence analyser

Info

Publication number: WO2003001190A1
Application number: PCT/FI2002/000562
Authority: WO
Inventors: Heikki SIPILÄ; Unto Paakkinen
Original assignee: Bs Instrument Technology Oy
Priority date: 2001-06-26
Filing date: 2002-06-26
Publication date: 2003-01-03
Also published as: FI20011362A0; EP1423683A1; FI20011362A

Abstract

An analyser arrangement is used for repeatedly analysing the composition of a target material flow. An X-ray radiation source (123, 205) irradiates the target material flow and a detector arrangement (120, 121, 303, 304, 305, 309) detects fluorescent radiation quanta. A cooler arrangement (308, 319, 500, 501, 502, 503, 504) keeps at least parts of the detector arrangement at a low temperature. Processing electronics (311, 401, 404, 405) convert detection results into concentration values that describe the analysed composition of the target material flow. The X-ray radiation source (123, 205) is an X-ray tube. The analyser arrangement further comprises a controllable high voltage source (204) that is coupled to said X-ray tube (123, 205). The detector arrangement comprises a solid-state semiconductor detector element (303), and the cooler arrangement (308, 319, 500, 501, 502, 503, 504) comprises a peltier element. The X-ray radiation source (123, 205), the high voltage source (204), the detector arrangement (120, 121, 303, 304, 305, 309), the cooler arrangement (308, 319, 500, 501, 502, 503, 504) and the processing electronics (311, 401, 404, 405) are all located within a single mechanical entity (107).

Description

X-ray fluorescence analyser, and a method for using an X-ray fluorescence analyser

The invention concerns generally the technology of analysing online the concentrations of certain desired substances in a target material flow. Especially the invention concerns the application of X-ray fluorescence detection in analysing the composition of the target material flow.

Online concentration analyses are of major importance in many industrial applications. As an example we may take the task of processing ore. It is very common to use a conveyor belt to move ore over relatively short distances in a mining and/or refining process. The ore transported by the conveyor belt constitutes a flow. It would be very useful to know, what useful minerals are present in the flow and in which proportional amounts at each desired time instant.

X-ray fluorescence is a known physical phenomenon in which incident X-ray radiation excites the atoms of a target substance. When an excited atom returns to a normal state, it emits a radiation quantum the energy of which has a value typical to the element in question. In a fluorescence analyser some of the emitted quanta are collected into a detector that is able to discriminate between quanta of different energies. From the relative numbers of detected quanta of different energies it is possible to estimate the relative proportions of different elements contained in the target substance.

From prior art there are known certain devices and methods for implementing X-ray fluorescence analysis in industrial applications such as the ore processing example given above. Apparatuses and services of this kind are manufactured and marketed by known companies like Gamma-Metrics, Amdel, Asoma and Quarcon. A granted Finnish patent number FI 97647 exists where a certain analyser is disclosed together with certain measurements and processing that aim at improving the reliability of the measurement results. However, the prior art apparatuses known at the priority date of this patent application do not come without drawbacks. These are mainly related to large size, complicated overall structure, long installation time and the use of hazardous radiation sources containing inherently radioactive substances. A particular drawback is related to the use of gas-filled detectors: their resolution is in the order of 18% of measured absolute energy value, which is too coarse a resolution for reliably differentiating between elements that are close to each other in the peri- odic system, iron and chromium being an exemplary pair. Another known drawback of prior art X-ray fluorescence analysers is their dependence on awkward external cooling arrangements. Liquid nitrogen must be regularly supplied in large storage tanks to keep certain detector types cool enough, and many commercially available analysers require connections to water mains and sewers for providing constant water cooling.

The poor resolution of gas-filled detectors has generated the need for continuous calibration of the analyser arrangements through laboratory testing of samples and subsequent application of computational regression analysis to the results given by the fluorescence analyser. Not only is the task of continuously analysing samples in laboratory tedious, but there are also remarkable problems related to taking these samples from the material flow. A sample should always be a good representative of a large amount of material subjected to fluorescence analysis, but pure chance may cause even large deviations from such representativeness.

It is an object of the present invention to provide an analyser arrangement for analysing the element composition of material flows, which analyser arrangement is suitable for use in a variety of industrial applications. It is a further object of the in- vention to provide such an analyser arrangement that is compact and user-friendly. Yet another object of the invention is to provide a method for using such an analyser arrangement effectively and accurately with less rigorous requirements for constant calibrating than in prior art arrangements.

The objects of the invention are achieved by placing the X-ray source, X-ray- detector, processing electronics and cooling arrangements of an analyser apparatus together into one, compact unit. Certain objects of the invention are also achieved by using a solid-state semiconductor detector for collecting and detecting the fluorescence quanta. Certain objects are also achieved by building the analysis algo- rithms on the basis of the so-called Blol iin formula that relates measured spectral line intensities with element concentrations in a measured sample.

The analyser arrangement according to the invention is characterized by the features that are recited in the characterizing part of the independent patent claim directed to an analyser arrangement. The invention concerns also a method that is characterized by the features that are recited in the characterizing part of the independent patent claim directed to a method.

According to a first aspect of the invention there is provided a compact analyser device in which an essentially single mechanical unit comprises an X-ray source and a detector as well as certain processing electronics and cooling arrangements. The device is self-contained to the extent that as contacts to the outside world there are preferably only a mains cable to be coupled to an electricity supply line as well as a data cable through which analysis results can be transmitted to a controlling computer and control commands can be received. In order to facilitate compactness and user-friendliness it is preferable to use a small X-ray tube as the radiation source instead of any inherently radioactive substances. Similarly in order to facilitate compactness it is preferable to use a solid-state semiconductor detector with peltier cooling.

According to a second aspect of the invention the processing electronics that are part of the compact analyser device are arranged to derive the analysis results from measured spectral lines through repeated application of a fundamental parameter model that is based on the so-called Bloldiin formula. This means that the measured spectral line intensities are placed into a group of equations which is solved numerically to obtain concentration values that should give rise to such measured spectral line intensities. The computed concentration values are normalized so that their sum equals unity, and suitable calculational methods are applied, repeatedly is necessary, so that the concentrations converge towards an optimal estimate.

The major advantage of using the above-described calculational method for obtaining the concentration values is that the analysis can be made almost completely independent of calibration, at least independent of such constant and repeated sam- pling and calibration that is typical to processes based on computational regression analysis. The analyser only needs some basic calibration that can most advantageously be performed at the manufacturer's premises as a part of set-up testing. Nevertheless, according to a third aspect of the invention it has been found advantageous to build the installation arrangement of the analyser device so that the device can be lifted and/or otherwise moved away from the usual measurement position, so that a calibration sample holder can be placed next to the X-ray tube and detector apertures. A calibration sample can then be placed in the holder so that the device measures the fluorescent characteristics of the calibration sample in the X-ray wave- lengths and uses the results to tune the values of certain parameters that form a part of the calculational process of obtaining concentration values.

The novel features which are considered as characteristic of the invention are set forth in particular in the appended claims. The invention itself, however, both as to its construction and its method of operation, together with additional objects and advantages thereof, will be best understood from the following description of specific embodiments when read in connection with the accompanying drawings.

Figs, la to Id illustrate an analyser device according to an embodiment of the invention in operating position over a conveyor belt,

Fig. 2 illustrates a radiation generator part of an analyser device according to an embodiment of the invention,

Fig. 3 illustrates a radiation detection part of an analyser device according to an embodiment of the invention,

Fig. 4 illustrates a processing part of an analyser device according to an embodiment of the invention,

Fig. 5 illustrates a cooler part of an analyser device according to an embodiment of the invention,

Fig. 6 illustrates an energy calibration method according to an embodiment of the invention,

Fig. 7 illustrates a primary calibration method according to an embodiment of the invention,

Fig. 8 illustrates a measurement method according to an embodiment of the invention and

Fig. 9 illustrates an analyser device according to an embodiment of the invention in an on-site calibration position.

Figs la, lb and lc are three mutually orthogonal views of an analyser device ac- cording to an embodiment of the invention in operating position over a conveyor belt. A basic support structure 101 and a number of rollers 102 attached thereto support a conveyor belt 103 (which for clarity reasons is not shown in Fig. la). The target material the composition of which should be analysed lies within the trough defined by the conveyor belt 103 and is shown cross-hatched in Figs, lb and lc. The support frame for the analyser device is built on struts 104 attached to the basic support structure 101 of the conveyor belt. Elastic vibration attenuators 105 are preferably used to attach a frame structure 106 to the struts 104, and the analyser device itself 107 is attached to the frame structure 106. The analyser device is represented in figs, la, lb and lc as a single mechanical unit 107, which according to an important principle of the invention houses a radiation source (an X-ray tube), a solid-state semiconductor detector, processing and analy- sis electronics as well as certain cooling arrangements. Later we will describe in more detail certain advantageous ways of arranging the electromechanical interconnections of the various parts of the analyser device.

The mechanical features shown in figs, la, lb and lc are exemplary, but they have their contributions to the advantageousness of the analyser device according to the invention when compared to many prior art devices. Using an X-ray tube as the radiation source means that the generation of ionizing, energetic electromagnetic radiation can be switched off whenever wanted. Heavy shielding against stray radiation is thus not needed, because radiation can always be switched off before human beings and/or devices sensitive to radiation come close to the device. The use of a (peltier-cooled) sensitive solid-state semiconductor detector means also that the required radiation intensities from the X-ray tube are relatively low, which further serves to eliminate the need for heavy shielding. Placing the analyser device above the conveyor belt allows the X-ray tube and detector apertures to be placed rela- tively near to the target material flow, while simultaneously reducing the risk of the analyser device getting dirty or damaged through material particles dropping and/or being thrown around. Elastic vibration attenuators reduce the amount of mechanical vibration coupled to the analyser device, which improves reliability. Simple mounting on struts allows the analyser device to be installed nearly anywhere where a conveyor belt exists. Placing the detector 108 "downstream", i.e. to that end of the analyser which does not meet first the material coming on the conveyor belt, reduces the risk of the detector or the nearby-located X-ray tube getting damaged by unexpectedly large material particles that reach higher than the usual material level.

In figs, lb and lc one may note that the analyser device has not been installed symmetrically above the center line of the conveyor belt. However, the installation places the detector aperture 108 directly above the center line of the conveyor belt, thus ensuring optimal geometric coverage for the detector. The arrangement of figs, la, lb and lc suggests that the detector's principal reception direction is vertical, or at least at a right angle against the moving direction of the conveyor belt (which in figs, la and lc is from right to left) while the X-rays are emitted from an output aperture of the X-ray tube obliquely so that the center point of the conical X-ray beam coincides with the point where the detector's main optical axis meets the target ma- terial. Fig. Id clarifies further this arrangement. The detector consists mechanically of a smaller housing 120 for the detector diode and a larger housing 121 attached thereto for a detection unit input stage. The central axis of the detector coincides with a vertical line 122 over the middle point of the conveyor belt. The optics of the detector and the X-ray tube 123 are preferably built so that the acceptance cone 124 from which the detector accepts incoming quanta is remarkably wider than the illumination cone 125 of the X-ray tube. This way it is ensured that despite of the oblique illumination angle of the X-ray tube the portion of target material irradiated with X-rays is always completely within the acceptance cone 124. This is advanta- geous, because it would be useless to waste power in producing X-rays that induce fluorescence at a location from which the fluorescent quanta can never reach the detector. Additionally a narrow X-ray beam is easiest to control so that radiation hazards remain at minimum. The widths of the illumination and acceptance cones means the value of their top angle in spatial angular units.

Fig. 2 is a schematic block diagram of what constitutes the so-called radiation generator part 200 of an analyser device according to an embodiment of the invention. A power supply line 201 from an electric power source of 85-260 VAC is coupled to a local supply voltage source 202 the task of which is to generate and regulate the internal supply voltages needed within the radiation generator part 200. From the local supply voltage source 202 there is a connection to a local check and control block 203 that operates essentially as a controllable switch supplying power to a high-voltage generator block 204 when needed. The last-mentioned generates and regulates the high voltage (in the order of several tens of kV) needed to operate an X-ray radiation tube 205. An indication circuit 206, equipped with an indicator lamp 207, is also present. The task of the indication circuit 206 is to lit up the indicator lamp 207 as a sign of activated X-ray radiation. There is also an external indicator lamp connection 208 from the indication circuit 206 in order to similarly drive a remote indicator lamp for example in the immediate vicinity of the analyser de- vice, at a security gate controlling the access to the analyser device and/or in a control room. Couplings 209, 210 and 211 to and/or from the check and control block 203, the local supply voltage source 202 and the indication circuit 206 respectively are provided in order to couple these to a central control block of the analyser device.

Fig. 3 is a schematic block diagram of what constitutes the so-called radiation detection part 300 of an analyser device according to an embodiment of the invention. A high voltage supply line 301 is coupled through an HV supply filter 302 to a de- tector diode 303, which is further coupled to a detection unit input stage 304. Physically the detector diode 303 is located next to a window 305 that allows X-ray radiation from outside to reach the detector diode 303 easily. There is also a check signal input line 306 that is coupled through a check signal input filter 307 to a point on the high voltage supply line 301 between the HV supply filter 302 and the detector diode 303. The detector diode 303 and the detection unit input stage 304 are located within a cooled container 308, which is further confined within a vacuum container 309 together with a magnetic vacuum pump 310. An output of the detection unit input stage 304 is coupled to a preamplifier 311, from which there is a further output 312 for conducting signals to further processing. There is also a supply voltage input line 313 to the preamplifier 311. Another component of the radiation detection part 300 is a temperature check circuit 314 that is coupled to receive an operating voltage through the preamplifier 311, to monitor the temperature of the vacuum container 309 with a suitable sensor 315 and to provide temperature indications through a temperature indication output line 316. Fig. 3 also shows the necessary supply voltage input lines 317 and 318 for the detection unit input stage 304 and the magnetic vacuum pump 310 respectively, as well as a pair of liquid coolant input and output lines 319 that transport liquid coolant to and from the detection unit input stage 304.

Fig. 4 is a schematic block diagram of what constitutes the so-called processing part 400 of an analyser device according to an embodiment of the invention. The heart of the processing part 400 is a programmable microprocessor device 401 the task of which is basically to control the operation of the other components of the analyser device. Bidirectional connections exist from the microprocessor device 401 to the other components of the processing part 400, which are a basic power supply 402, a high- voltage power supply 403, an amplification, filtration and A/D-conversion block 404, a windowed pulse count block 405 and a check, control and indication panel 406. In addition of these there is an RS-485 type serial connection interface adapter 407 with which the microprocessor device 401 also has a bidirectional connection. Other connections to and/or from the microprocessor device 401 include an RS-232 type serial connection 408 to the external world, a bidirectional control connection 209 to and from the local check and control block 203 of the radiation generator part 200 (see fig. 2), a bidirectional control connection 210 to and from the local supply voltage source 202 of the radiation generator part 200 (see fig. 2), a unidirectional control connection 211 to the indication circuit 206 of the radiation generator part 200 (see fig. 2), a unidirectional input connection 316 from the temperature check circuit 314 of the radiation detection part 300 (see fig. 3) and a uni- directional control connection 409 to a coolant pump control circuit in a cooler part that is to be described with reference to fig. 5. From the serial connection interface adapter 407 there is a bidirectional connection 410 to the external world; this counts in fact also as one of the connections of the microprocessor device 401 because the role of the serial connection interface adapter 407 is only to make certain protocol conversions according to the RS-485 specifications.

The microprocessor device 401 receives its operational voltage from the basic power supply 402, which in turn receives an AC input voltage from the external world along a supply voltage line 411. From the basic power supply 402 there are, within the processing part 400, operational voltage connections also to other components that need operational voltages within "normal" (not high-voltage) levels. Additionally the basic power supply 402 supplies operational voltages to the radiation detection part 300: along a supply voltage line 313 to the preamplifier 311 and a supply voltage line 317 to the detection unit input stage 304 (see fig. 3). From the high- voltage power supply 403 there are high-voltage supply lines 301 and 318 to the detector high- voltage input and the magnetic vacuum pump 310 respectively of the radiation detection part 300 (see fig. 3).

The amplification, filtration and A/D-conversion block 404 receives signals along an input line 312 from the preamplifier 311 of the radiation detection part 300 (see fig. 3) and directs its output signals into the windowed pulse count block 405. From the latter there is a feedback connection back to the amplification, filtration and A/D-conversion block 404 as well as an output line 412 to the external world.

Fig. 5 is a schematic block diagram of what constitutes the so-called cooler part 500 of an analyser device according to an embodiment of the invention. A coolant pump 501 is coupled to circulate liquid coolant through a cooler unit 502; the input and output lines 319 for the coolant are coupled to the detection unit input stage 304 of the radiation detection part 300 (see fig. 3). A fan 503 can optionally be used to maintain a constant flow of fresh air around the cooler unit 502 so that the cooling effect thereof would be as good as possible. In very dusty conditions where it is preferable not to have any air intakes (not even filter-equipped ones) in the outer cover of the analyser device it is possible to omit the fan 503 and to place the cooler unit 502 so that convective cooling from outside of the analyser device takes care of transporting heat off the cooler unit 502. A control circuit 504 receives a supply voltage and control commands along lines 505 and 409 respectively. The latter is a unidirectional (or possibly bidirectional) connection (to and) from the microprocessor device 401 of the processing part 400 (see fig. 4).

In the embodiment of the invention illustrated in figs. 2 to 5 the cooling of the de- tector arrangement is implemented as a combination of peltier and liquid cooling. The cooled container 308 contains a peltier element that has a cold end and a hot end. Of these the cold end is arranged to keep the detector diode 303 and the charge-sensitive, pre-amplifying detection unit input stage 304 at a very low temperature, preferably in the order of -100°C. The hot end is in contact with a con- verter (not separately shown) where the liquid coolant that circulates through the lines 319 absorbs heat from the hot end of the peltier element.

In addition to a low temperature the optimal operation of the detection arrangement requires the immediately surrounding gas atmosphere to be as thin as possible. Therefore the cooled container 308 is further enclosed in the vacuum container 309 the inside of which is kept substantially empty of gasses with the magnetic vacuum pump 310. The magnetic vacuum pump 310 should be continuously operational, which means that the supply voltages arrangement of the analyser device should be arranged to provide the magnetic vacuum pump 310 with an operational voltage even during periods when the analyser device is otherwise turned off. For periods of transportation there should be provided a supplementary power source that keeps the magnetic vacuum pump 310 going continuously.

During operation the microprocessor device 401 instructs the local supply voltage source 202 and the local check and control block 203 of the radiation generator part 200 to use the high-voltage generator block 204 for supplying the X-ray radiation tube 205 with the necessary voltage(s) so that it emits X-ray radiation into a target. The microprocessor device 401 also instructs the high- voltage power supply 403 to provide a high voltage through the HV filter 302 to the detector diode 303. Addi- tionally the microprocessor device 401 uses its control connections to other components of the analyser device to establish an operational state where each component receives its necessary operational voltage(s) from the appropriate voltage source(s), and the indication circuit 206 provides a warning about the X-ray radiation being active. As a part of the operational state the cooling arrangement performs its task of keeping the detector arrangement at a reasonably low temperature. The temperature monitoring system that relies on blocks 314 and 315 may be used to take into account the temperature dependence of certain operational parameters and/or to simply provide the microprocessor device 401 with information about whether or not the detector arrangement is rurining cool enough.

Fluorescent radiation quanta from the target come through the window 305 and hit the detector diode 303 causing charge pulses through the diode. According to the known principle of fluorescence analysis the charge represented by each pulse is proportional to the energy of the quantum that caused the pulse. These pulses are gathered and the energies represented thereby are measured in the signal processing chain consisting of the detection unit input stage 304, the preamplifier 311 and the amplification, filtration and A/D-conversion block 404. The windowed pulse count block 405 composes output information that tells, how many pulses of which energy were received in each time window of predefined length. This information constitutes the obtained measurement spectra and can be used as such for further processing, if it is taken out through the output line 412. On the other hand the microproc- essor device 401 can be used to perform calculations on the basis of the obtained measurement spectra to reveal the element concentrations within the target in a form suitable for storing and/or further processing.

The functional blocks described above with reference to figs. 2 to 5 should all be placed within the single mechanical unit 107 shown in figs, la to lc in order to achieve the advantageous feature of unprecedented compactness. Connections to the outside world, that means lines 201, 208, 408, 410, 411, 412 and 505 can be taken to and from said single mechanical unit 107 to a control room or suchlike through suitable cabling. The check signal input 306 shown in fig. 3 is preferably just a con- nector within the analyser device with no external connections, because it is mainly used at the manufacturing stage for inputting test impulses to the detector arrangement. If needed, a cable can be used to couple also the check signal input 306 to some kind of control room arrangements that can then be used to generate test impulses during operation-time testing.

Next we will describe the advantageous method that the analyser device according to the invention is programmed to implement in calculating the element concentrations. Let us assume that monochromatic excitation radiation with a certain energy εO has been used to obtain a measurement spectrum, in which there are spectral lines the measured intensities of which are known. There is known the so-called Blokhin formula '.A - ',fik,r_Λ,F,c, ' - ^« < •> * * ^ft lr ₍₁₎

where I_ε , = the measured intensity of the spectral line with energy ε and coming from the :th element, A, = a correction coefficient for the /^':th element, close to 1

I_ε0 = the intensity of the incident radiation with energy εO,

G = a scalar geometric factor dependent on distances and angles, k = a constant for the /:th element, depending on atomic mass r_-0 , = mass absorption coefficient for the incident radiation with en- ergy εO and the z:th element,

F_ε ⁼ apparatus response function for the energy ε, C, = concentration of the z:th element, μ_ε0 = mass absorption coefficient in the sample for the incident radiation μ_ε = mass absorption coefficient in the sample for radiation with energy ε, φ - the angle between incident radiation and local horizontal in radians, ψ = the angle between detected radiation and local horizontal in radians,

M_s = surface density, calculated as M_s = pd , where p is the mass density of the sample and d is its thickness and T, = part of secondary fluorescence in the spectral line.

Of the mass absorption coefficients one should note that

M_to = ∑τ_e0ιJC_J (2)

and

where τ_{ε j} is the mass absorption coefficient for radiation with energy £- and they:th element. The r-values are known as such and can be read from commonly known tables. Regarding the apparatus response function F_ε it suffices to note that it represents the probability of a given quantum of energy ε being detected. It is close to 1 for quantum energies between 5 and 50 keV, but becomes lower for quanta with lower or higher energies, because the former may be absorbed in the input window and the latter may pass all the way through the detector diode without causing interactions. If the sample to be measured is thick, its surface density M_s comes big enough for approximating the formula (1) with a simpler one

I_ε Λ = ^I _*oGk_lτ_εQιlF_εC₁ (4) μ_ε ύnφ + μ_εlsmψ

which can be solved for the concentration to obtain

so in order to find the concentrations is should suffice simply to know the mass absorption coefficients. However, the latter are also functions of the concentrations, so it becomes impossible to solve the problem for the concentrations in closed form.

In the method according to the invention there is applied an iterative calculation scheme where one performs the steps of: a) taking first a coarse assumption of the element concentrations, b) if the sum of coarsely assumed concentrations is not equal to one hundred per cent, normalizing these so that their sum becomes equal to one hundred per cent, c) using the normalized assumed element concentrations in the mass absorption coefficients of formula (5) (or formula (1)) to obtain a set of refined mass absorption coefficients, d) normalizing the refined mass absorption coefficients so that their sum becomes equal to one hundred per cent and e) repeating steps c) and d) until a certain termination condition is fulfilled.

A reasonably good coarse assumption for the purposes mentioned above can usually be obtained by assuming that the concentrations are proportional to the measured spectral line intensities (or corrected spectral line intensities, which as a concept will be described later). Another way of obtaining reasonably good coarse assumptions is to use the calculated concentration values of a certain earlier measurement, most preferably either those obtained from an immediately previous measurement or a mean value of those obtained from X previous measurements where X is an integer. The termination condition can be as simple as always performing a constant number of iterations. More elaborate termination conditions can be devised from the principle that iteration is allowed to terminate when the change in concentration values between successive iteration rounds becomes smaller than a predetermined limit. Conditions of the latter kind are usually based on calculating the sum of squares of differences between respective concentration values and comparing it to a predetermined threshold value.

The above-presented example was based on the assumption that the excitation radiation is monochromatic, which is not the case when one uses simply an X-ray tube as a radiation source. The continuous range of energies emitted by the X-ray tube can be taken into account by replacing the simple multiplicative intensity term I_ε0 with an energy integral, which e.g. in the case of formula (4) means a modification into the form

i. A = G*ΛC, J ^J^ dE o, μ_ε0/ ,sm. φ + μ_ε/ ,sm. φ r, (6)

where — — is the intensity of exitation radiation within a given energy interval dE. dE

If one wants to account fully for the properties of the X-ray tube, one should sum the integral over bremsstrahlung radiation with a sum over the spectral lines of the X-ray tube.

There remains the question of how should one obtain values for the fixed parameters and other essentially constant factors in the calculation formulas. According to the invention this is accomplished during a procedure of calibration, where the analyser device measures standard samples the compositions of which are known.

Calibration is divided here into energy calibration, primary calibration and coefficient calibration. The purpose of the first of these is to establish a truthful relationship between the channel numbers in the analyser device and the energies of the received quanta. As is widely known, digital spectral analysers take each received and measured quantum into account by representing it as an entry in a channel or bin that covers that part of the energy axis where the measured energy of the quantum is found. The channels have ordinal numbers, and the relationship between the energy ε of a quantum and the number n of the channel it goes into should be of a linear form

ε = An + B (7)

where A and B are real constants. Energy calibration means finding the exact values of A and B. It is most straightforwardly implemented so that a few samples of known concentration are analysed and the known spectral lines obtained therethrough are associated with their respective channel numbers. As an example when one has analysed elements X and Y, one has obtained a spectrum where the K^ peak of element X is centered at channel number n_x and the K_α peak of element Y is centered at channel number riγ. The characteristic energies ε^_a and εχ_aγ respectively of these peaks are known, so one simply solves A and B from the group of equations

The reliability of energy calibration can be improved by measuring more spectral lines than two, replacing the group of equations (8) with another group of equations where each measured spectral line appears as an equation of its own, and using a suitable line fitting algorithm to find the most appropriate values for A and B.

Fig. 6 is a schematic method diagram of the above-explained energy calibration method. At step 601 a user instructs the analyser device to start acquiring an energy calibration spectrum; before that the user must have placed an energy calibration sample so that the analyser device can analyse it. At step 602 the analyser device acquires an energy calibration spectrum and at step 603 it displays it to the user e.g. on a display screen of a computer that is coupled to the analyser device through one of the RS-232 or RS-485 serial interfaces. At step 604 the user selects a spectrum peak he believes to have reliably recognised by e.g. clicking it with a mouse. The selection is conveyed to the analyser device, which notes the channel number that corresponds to the center of the indicated spectrum peak. At step 605 the analyser device prompts the user to enter the exact energy value that should correspond to the indicated peak. At step 606 the user gives the energy value. Steps 604-606 are repeated for a desired number of times, after which at step 607 the analyser device uses the obtained correspondences between channel numbers and energy values to calculate and store the constants it needs for unequivocally associating arbitrary channel numbers with energy values. Primary calibration can be further subdivided into spectral line intensity correction and the calculation of the constant parameters. The first of these means taking into account certain factors that make the measured intensity of a spectral line differ from an ideally obtained intensity, and the second means obtaining and fme-tuning the essentially constant parameter values that appear in formulas like (1), (5) and (6) in addition to the actual unknowns that are the concentrations and mass absorption coefficients.

All steps of acquiring a measurement spectrum with an analyser device according to the invention should include subtraction of background radiation as well as some kind of integration over time in order to lessen the effect of count rate fluctuations that are only caused by the stochastic nature of the processes involved in X-ray radiation and detection. Subtraction of background radiation and integration over time also form a part of spectral line intensity correction, but because they can be implemented in a known way using non-parametric linear low pass filters, they are not described here in detail.

A more important part of spectral line intensity correction is the correction for es- cape peaks, overlapping peaks and double peaks. As an example we may discuss the case of iron that reveals itself in an X-ray fluorescence spectrogram by giving rise to two nearly overlapping peaks, the Fe-K_α peak and the Fe-K_β peak. Separating these from each other can be accomplished with known least-sum-of squares methods. After such a separation has been made from a calibration spectrum that is based on a long measurement time and therefore has high statistical reliability, and the initial and final spectral line images have been stored for later reference, the analyser can apply a similar process in scaled form to convert e.g. any obtained Fe-K_α peak measurement that includes partial overlap from a nearby Fe-Kp peak into a pure Fe- K_α peak spectral line image.

Escape peaks appear in a measured spectrum because following a photoelectric absorption event near the surface of the detector diode the fluorescence photon generated in the photoelectric absorption event may escape the detector without being re- absorbed. The calibration process should include a step where the intensity repre- sented by the escape peak(s) is added to the measured intensity of the actual main peak. Similarly double peaks should be accounted for: they are a result of two radiation quanta arriving essentially simultaneously into the detector, so the intensity represented by the double peaks should also be added to the measured intensity of the actual main peak. A yet further step of spectral line intensity correction is that of replacing a measured non-Gaussian peak profile (caused by partial charge loss in the detector) with an appropriate Gaussian one. Such replacing can be accomplished e.g. by using some well-known curve-fitting algorithm to obtain a best possible fit between a Gaussian curve and the measured non-Gaussian peak profile.

It is common to all aspects of spectral line intensity correction that after they have been initialized with reference to a calibration spectrum that is based on a long measurement time and therefore has high statistical reliability, they provide a stored basis for replacing a non-ideal measured spectrum with a set of calculationally corrected peak profiles that together form the line structure of the spectrum. The intensities represented by such calculationally corrected peak profiles can then be used as inputs to formulas like (1), (5) or (6) when one tries to obtain the correct concentration values.

The calculation of the constant parameters in the primary calibration phase is most advantageously accomplished so that one lets the analyser device analyse a number of pure element samples having different atomic numbers. One series of pure element samples that has been found advantageous in primary calibration is such where the first sample is Si, the second sample is Ca, the third sample is Ti, the fourth sample is Fe, the fifth sample is Cu, the sixth sample is Ge, the seventh sample is Mo, the eighth sample is In or Sn, the ninth sample is W and the tenth sample is Pb or Bi. Assuming that energy calibration and spectral line intensity correction have been made, analysing each sample produces an energy calibrated and spectral line intensity corrected spectrum. Because each sample is pure, in the formulas for concentration calculation the mass absorption coefficients are reduced from sums to single terms, and the concentration terms become equal to unity. Gathering a num- ber of measured, energy calibrated and spectral line intensity corrected spectra with different pure element samples allows a group of equations to be written so that the most appropriate values for the constant coefficients can be solved from the group of equations either analytically in closed form or, more typically, with numerical, approximative methods that are known as such to a person skilled in the art.

The series of pure element samples suggested above has the advantageous feature of covering essentially the whole useful range of elements in the periodic table with nearly constant intervals between successive sample elements. Interpolation can be used to obtain the element-specific parameter values for elements that are not included in the series of pure element samples but lie between certain two successive sample elements in the periodic table. As a final part of primary calibration there are stored the obtained constant parameter values that can later be used as such in the formulas for concentration calculation. The process of primary calibration may include also steps where the values of certain constant parameters are entered manu- ally as axiomatic facts. Such manually entered parameter values may include e.g. measured values such as the mean distance between the target flow to be analysed and the detector, detector thickness, and correction factors for absorption in the detector window.

As a part of the information to be stored there are advantageously stored also a list of the elements and/or molecular compounds the concentrations of which should be analysed. As an example we present a case where there are stored two lists of the following kind:

The detectable element descriptions in the sample {ele- ment~its_atomic_shell}={Ca~K}; {Ti~K}; {Cr~K}; {Mn~K}; {Fe~K}; {Ni~K};

{Cu~K}; {Ta~L}; {W~L}; {Zn~K}; {Pb~L}; {Bi~L}; {Sn~K}; {Zr~K}; {Nb~K};

{Mo~K}; {Rh~K}; {In~K}; {Ge~K};

The description for possible chemical formulas in the class of samples {for- mula~molecular_concentration}={CaCO3~0}; {TiO~0}; {Cr2O3~0}; {MnO~0};

{Fe2O3~0}; {Ni~0}; {Cr~0}; {Sr~0}; {H20-2.5}; {MgO~10}; {Si02~14.7};

{A12O3-2.4};

The first list includes all those pure elements that the analyser device should be able to recognize either during primary calibration or during operative use. Such a list is useful in the sense that since the energies of all listed spectral lines all known beforehand, during primary calibration or operative use the analyser device will only try to locate and measure spectral lines that it finds included in this limited set. The second list is even more useful during operative use, since it lists those elements and chemical compounds that are known to be present in the flow of target material to be analysed and equally to exclude measured spectral lines that are known to come from an error source. In the practical example given above the second list does not contain zinc, because the ore to be analysed is known not to contain zinc in any significant amounts while the conveyor belt used to transport the ore under the ana- lyser device is known to contain zinc. Not including zinc in the second list means that the analyser device is instructed to neglect any zinc-specific spectral lines because they do not come from the actual measured target material flow. In the second list there are two types of entries: those where the tilde is followed by a zero and those that have a numerical value after the tilde. This is a notational practice used to denote the fact that for elements and compounds of the first kind the concentration is unknown from the beginning and should be assumed from meas- urement results, while for elements and compounds of the second kind the coarsely assumed concentration is the one that appears after the tilde. In ore-processing applications there are usually certain basic compounds that are always present in the ore in almost equal percentage, so giving this assumed percentage as initial information to the analyser device may considerably speed up the process of convergently calculating the concentrations of the "interesting" elements and compounds.

The fact that certain elements are given on the second list only in molecular compounds reflects the fact that also in the ore they are not present as pure elements but in these molecular compounds. Announcing the compounds means that the analyser device has the possibility of analysing and announcing also the concentrations of light elements such as H, C or O that are otherwise invisible in X-ray fluorescence spectroscopy. One should note that certain compounds that appear in the above- given exemplary list are completely invisible in X-ray fluorescence spectroscopy, which means that if they were present in the target flow and were not accounted for by giving their assumed approximate concentrations in the list, the analyser device would completely neglect them in announcing the calculated concentrations for other elements and compounds, thus greatly exaggerating them from their actual values.

The use of the second list allows the analyser device to be easily adjusted to a new kind of ore or other target flow material. Complete recalibration is not needed; one simply takes representative samples from the new target flow material, analyses them in laboratory to find out the concentrations to be "preprogrammed", and stores these as a part of a new second list after which the analyser device can quickly be put into operation again.

The analyser device may even be allowed to edit the second list itself. This means that the assumed approximate concentrations in the list may be updated to reflect e.g. certain long-term mean values of calculationally obtained concentrations.

There remains the third kind of calibration mentioned previously, namely coefficient calibration. In the analyser device according to the invention this means simply that practice has shown there to be an almost linear dependence between con- centration values obtained with the calculational methods described above and real concentration values obtained in laboratory measurements. The invention provides for a possibility of storing two real parameters a and b so that a final concentration value y is calculated from a concentration value x obtained with the calculational methods described above by letting y = ax + b. The default values of a and b should be 1 and 0 respectively, meaning that the calculational methods give exact values. However, this is not always true, so coefficient calibration may be performed by taking several spectra of a real target material flow, calculating the concentrations x, comparing these with laboratory-analysed concentrations y taken from representa- tive samples of the same real target material flow and using a line-fitting method to find out appropriate values for the coefficients a and b. These can be stored in the analyser device, which can then use them to proactively convert further otherwise calculationally obtained concentration values x to assumedly more correct values y.

In every case where the analyser device is said to store some kind of information it is assumed that suitable storing means exist within the appropriate components. Storing means are not explicitly shown in the drawings, because they constitute a routine part of components like microprocessor devices, amplification, filtration and A/D-conversion blocks as well as windowed pulse count blocks.

Fig. 7 is a schematic flow diagram that illustrates primary calibration according to an advantageous embodiment of the invention. At step 701 the analyser device acquires a measured spectrum of a pure element sample. Steps 702 and 703 represent the correction made for escape peaks, steps 704 and 705 represent the correction made for double peaks, steps 706 and 707 represent the correction made for non- gaussian measured peak profile, and steps 708 and 709 represent the correction made for overlapping peaks. At step 710 the analyser device compares the original and corrected form of each processed peak and stores a correction function that can later be used for mapping any obtained measurement result of a peak having the same energy into a more ideal peak for calculating the concentrations according to the fundamental parameter model.

Fig. 8 is a schematic flow diagram that illustrates operative use of an analyser device according to the invention. At step 801 the analyser device acquires a measured spectrum of a target by irradiating it with X-rays and detecting fluorescent quanta. The time interval for obtaining a single measured spectrum can be selected according to conditions like speed of the conveyor belt (if one is used) and assumed frequency of remarkable changes in concentrations in the target material flow. At step 802 the analyser device locates a peak in the acquired spectrum and recognises it by comparing the energy represented by the peak to the known energies of those peaks that it has been instructed to look for. At step 803 the analyser device corrects the intensity of the located and recognised peak by utilizing a correction function stored at step 710 of primary calibration. At step 804 it stores the corrected intensity of the peak. Step 805 is a check for any remaining unprocessed peaks in the spectrum; a positive finding at step 805 causes a jump back to step 802 while a negative finding at step 805 allows proceeding into step 806, where default concentrations for "invisible" elements and/or compounds are read from memory. At step 807 the ana- lyser device converts the stored peak intensities into coarse assumptions of concentrations and normalizes these appropriately.

At step 808 the analyser device derives estimated mass absorption coefficients from the currently available concentration values and uses them to obtain a next round of concentration values, which in turn are normalized at step 809. Step 810 is a check for whether the termination condition for iterative refining of the concentration values has been met; a negative finding at step 810 causes a jump back to step 808 while a positive finding at step 810 allows proceeding into step 811 where the results of the iterative calculation are presented. Linear transforming of the results ac- cording to what was explained in association with coefficient calibration is not separately shown in fig. 8, but it can easily be combined with the output step 811.

The results of one run through the method of fig. 8 only reveal the concentrations in that portion of the target material flow that was subjected to irradiation and detec- tion at step 801. For practical usability of the analysis results it is advantageous to keep the analysing process going on continuously, so that the operative time is divided into successive time windows, a concentrations result is calculated for each time window separately and results obtained in successive time windows are combined in a variety of statistical ways. Non-limiting examples of such combining are the calculation of running mean values over a number of latest time windows as well as the calculation of mean values and variations per each hour, workshift or day.

It is important to note that energy calibration and primary calibration can and should be performed as a part of the final operational tests of a newly manufactured analyser device before even delivering it to the site of operational use. It has been reported that an analyser device according to the invention has been operational and giving meaningful analysis results only four hours after delivery to the site of opera- tional use, which is a remarkable improvement over prior art analyser solutions that might have required weeks or months of installing, on-site testing and calibration as well as continuous laboratory analyses for the purposes of computational regression analysis.

If however one wants to perform calibration at the site of operational use, there should be an easy way for holding known samples in front of the X-ray tube and detector parts of the analyser device. Fig. 9 illustrates a modification of the simple installation of fig. lb that allows a calibration sample holder 901 to be detachably at- tached to the underside of the analyser device. The calibration sample holder 901 is a kind of flat spoon. The analyser device body 107 or some of the associated installation structures comprises a socket for receiving and holding one end of the handle of the calibration sample holder 901. A calibration sample 902 is thereby easily held in position in front of the X-ray tube and the detector. In order to ensure that there is enough space to use the calibration sample holder 901, two of the struts 104 are equipped with rotational joints 903 while the other two are equipped with detachable joints. The latter are detached and the whole analyser device assembly with its main body 107 and frame structure 106 is lifted upwards so that it turns around the common rotational axis of the rotational joints 903. A supporting arm 904 is most advantageously used to hold the analyser device assembly in place while it is in its elevated position.

The exemplary embodiments of the invention presented in this patent application are not to be interpreted to pose limitations to the applicability of the appended claims. The verb "to comprise" is used in this patent application as an open limitation that does not exclude the existence of also unrecited features. The features recited in depending claims are mutually freely combinable unless otherwise explicitly stated.

Claims

1. An analyser arrangement for repeatedly analysing the composition of a target material flow, comprising:

- an X-ray tube (123, 205) for irradiating the target material flow with fluorescence- inducing X-rays,

- a controllable high voltage source (204) coupled to said X-ray tube (123, 205),

- a detector arrangement (120, 121, 303, 304, 305, 309) that comprises a solid-state semiconductor detector element (303) for detecting fluorescent radiation quanta emitted from the irradiated target material flow, - a cooler arrangement (308, 319, 500, 501, 502, 503, 504) that comprises a peltier element for keeping at least parts of the detector arrangement at a low temperature and

- processing electronics (311, 401, 404, 405) that are arranged to convert detection results generated within the detector arrangement into concentration values that de- scribe the analysed composition of the target material flow; characterized in that:

- the X-ray tube (123, 205), the high voltage source (204), the detector arrangement (120, 121, 303, 304, 305, 309), the cooler arrangement (308, 319, 500, 501, 502, 503, 504) and the processing electronics (311, 401, 404, 405) are all located within a single mechanical entity (107) and

- the processing electronics (311, 401, 404, 405) are arranged to apply a stored list of the elements and/or molecular compounds the concentrations of which should be analysed.

2. An analyser arrangement according to claim 1, characterized in that said single mechanical entity comprises an essentially closed container (107) for enclosing the X-ray tube (123, 205), the high voltage source (204), the detector arrangement (120, 121, 303, 304, 305, 309), the cooler arrangement (308, 319, 500, 501, 502, 503, 504) and the processing electronics (311, 401, 404, 405); and the analyser ar- rangement further comprises a number of support struts (104) for supporting said essentially closed container over a target material flow.

3. An analyser arrangement according to claim 2, characterized in that

- said essentially closed container (107) is elongated in a certain direction having a first end and a second end,

- said support struts (104) are arranged to support said essentially closed container (107) over the target material flow so that said first end faces the mcoming direction of the target material flow and said second end faces the outgoing direction of the target material flow, and

- the detector arrangement (108, 120, 121, 303, 304, 305, 309) is located within said essentially closed container (107) closer to said second end than said first end.

4. An analyser arrangement according to claim 2, characterized in that the detector arrangement (108, 120, 121, 303, 304, 305, 309) is located within said essentially closed container (107) so that its principal reception direction coincides with a line (122) that - is perpendicular to the direction of the material flow and

- in a direction across the target material flow intersects the target material flow in the middle thereof.

5. An analyser arrangement according to claim 4, characterized in that the X- ray tube (123, 205) is located within said essentially closed container (107) so that its principal irradiation direction is at an oblique angle against the principal reception direction of the detector arrangement (108, 120, 121, 303, 304, 305, 309).

6. An analyser arrangement according to claim 5, characterized in that - the X-ray tube (123, 205) is arranged to emit X-rays into an illumination cone (125) of a certain width,

- the detector arrangement (108, 120, 121, 303, 304, 305, 309) is arranged to accept fluorescent radiation quanta in an acceptance cone (124) having a width that is remarkably larger than the width of said illumination cone (125), and - said illumination cone (125) and said acceptance cone (124) are arranged to intersect so that within measurable parts of the target material flow said illumination cone (125) is completely within said acceptance cone (124).

7. An analyser arrangement according to claim 2, characterized in that - two of said support struts (104) comprise a rotational joint (903),

- the rotational joints (903) have a common rotational axis, and

- at least one other support strut (104) comprises a detachable joint; so that detaching said detachable joint allows the essentially closed container (107) to be moved around said rotational axis into a direction that is away from the target material flow.

8. An analyser arrangement according to claim 7, characterized in that it additionally comprises a calibration sample holder (901) and - within either said essen- tially closed container (107) or a support structure thereof (106) - a socket for receiving and holding one end of said calibration sample holder (901), so that when held by one end by said socket in a configuration where the essentially closed container (107) has been moved around said rotational axis into a direction that is away from the target material flow, said calibration sample holder (901) is arranged to hold a calibration sample (902) in front of the X-ray tube and the detector arrangement.

9. A method for repeatedly analysing the composition of a target material flow, comprising the steps of:

- irradiating the target material flow with fluorescence-inducing X-rays by generating a controllable high voltage and using it to accelerate electrons in an X-ray tube,

- detecting (801) fluorescent radiation quanta emitted from the irradiated target material flow by measuring the deposited charges associated with current pulses through a solid-state semiconductor detector element, said current pulses being induced by said radiation quanta,

- keeping at least parts of the arrangement that performs said detecting at a low temperature by conducting thermal energy from said solid-state semiconductor detector element into a cold end of a peltier element and - converting (802, 803, 804, 805, 806, 807, 808, 809, 810, 811) detection results generated within the detecting arrangement into concentration values that describe the analysed composition of the target material flow; characterized in that:

- the steps of generating a controllable high voltage and using it to accelerate elec- trons in an X-ray tube, measuring the deposited charges associated with current pulses through a solid-state semiconductor detector element and conducting thermal energy from said solid-state semiconductor detector element into a cold end of a peltier element, as well as the step of converting detection results generated within the detecting arrangement into concentration values, are all accomplished within a single mechanical entity and

- the step of converting detection results generated within the detecting arrangement into concentration values comprises a substep of applying a stored list of the elements and/or molecular compounds the concentrations of which should be analysed.

10. A method according to claim 9, characterized in that the step of converting detection results generated within the detecting arrangement into concentration values comprises the substeps of: a) providing (802, 803, 804, 805, 806, 807) a coarse assumption of concentration values, b) using said coarse assumption of concentration values to obtain (808) assumed mass absorption coefficients and c) using (808, 809) said assumed mass absorption coefficients to calculate refined concentration values.

11. A method according to claim 10, characterized in that steps a) and c) each comprise the substep of normalizing (807, 809) the concentration values concerned so that their sum becomes equal to a value representing full content.

12. A method according to claim 11, characterized in that after step c) it comprises the steps of d) using the refined and normalized concentration values to obtain (808) refined as- sumed mass absorption coefficients, e) using (808) said assumed mass absoφtion coefficients to calculate further refined concentration values, f) normalizing (809) the further refined concentration values so that their sum becomes equal to a value representing full content and g) repeating steps d), e) and f), always using the most refined concentration values and mass absorption coefficients available, until a certain terminating condition (810) has been met.

13. A method according to claim 10, characterized in that the step of providing a coarse assumption of concentration values comprises the substeps of:

- detecting (802) spectral line intensities for certain constituents of the target material flow,

- converting (803, 804, 805) the detected spectral line intensities into coarsely assumed concentration values according to the relative magnitudes of the detected spectral line intensities and

- reading (806) coarsely assumed concentration values for certain other constituents of the target material flow from a stored list.

14. A method according to claim 10, characterized in that the step of using said coarse assumption of concentration values to obtain (808) assumed mass absoφtion coefficients comprises application of the formulas μ_ε0 = ∑τ_{εo j}C_j and

^ = Σ^r,,;^C > ^Where τ_{ε0 ]} = mass absoφtion coefficient for fluorescence-inducing X-rays with energy εO and the :th element, τ_ε = mass absoφtion coefficient for radiation with energy ε and the

/:th element C - concentration of they :th element, μ_ε0 = mass absoφtion coefficient in the sample for the fluorescence-inducing X-rays and μ_ε = mass absoφtion coefficient in the sample for radiation with energy ε.

15. A method according to claim 10, characterized in that the step of using (808) said assumed mass absoφtion coefficients to calculate refined concentration values comprises the step of using the formula

where I_ε , = the measured intensity of the spectral line with energy ε and coming from the t:th element, 4 ⁼ a correction coefficient for the z:th element, close to 1 G = a scalar geometric factor dependent on distances and angles, k_l = a constant for the 7^':th element, depending on atomic mass

F_ε ⁼ apparatus response function for the energy ε,

C, = concentration of the i.th element,

— ^- = is the intensity of exitation radiation within a given energy in- dE terval dE, τ_ε0, = mass absoφtion coefficient for the incident radiation with energy εO and the /:th element, μ_εQ = mass absoφtion coefficient in the sample for the incident radiation μ_ε = mass absoφtion coefficient in the sample for radiation with energy ε, φ = the angle between incident radiation and local horizontal in radians, ψ = the angle between detected radiation and local horizontal in radians and r, = part of secondary fluorescence in the spectral line.

16. A method according to claim 10, characterized in that as a preparatory measure for enabling accurate analysis of the composition of a target material flow it comprises, before step a), a calibration process that comprises the steps of:

- initializing in a process of energy calibration (601, 602, 603, 604, 605, 605, 606, 607) the correspondence of channels in an analyser arrangement performing the analysis and energies of detected quanta,

- establishing spectral line intensity correction functions (701, 702, 703, 704, 705, 706, 707, 708, 709, 710, 711) for converting measured spectral line intensities into corrected spectral line intensities and

- deriving (712) values for calculational parameters that are essentially independent of concentrations from analysis results of samples with known composition.

17. A method according to claim 16, characterized in that it additionally comprises a further calibration step where a constant correction function is established between calculationally obtained analysis results and experimentally verified target flow compositions.

18. A method according to claim 16, characterized in that the step of establishing spectral line intensity correction functions comprises the substeps of:

- establishing an escape peaks correction (702, 703) that adds intensities represented by escape peaks into the intensities of main peaks,

- establishing an overlapping peaks correction (708, 709) that separates overlapping peaks from each other,

- establishing a double peaks correction (704, 705) that adds intensities represented by double peaks into the intensities of main peaks and

- establishing a Gaussian fit correction (706, 707) that replaces non-Gaussian peak profiles with Gaussian ones.