JPH0416728B2 - - Google Patents

Description

BACKGROUND OF THE INVENTION This invention uses a spectroscopic analyzer, and in particular a flash tube with high light intensity, to emit light.
This invention relates to an improved spectroscopic analysis system.

Xenon flash tubes are used in a variety of different types of equipment to provide short-duration flash beams of extremely high light intensity. Generally, a flash tube is a glass sealed tube into which a certain electrode is inserted and xenon gas is sealed inside. usually,
When an arc is generated between these electrodes, energized by the discharge of the capacitor, and impinges on the pole face, a flash of light with high light intensity is generated within the tube.

In typical applications, the life of a xenon tube is determined by erosion of the arc electrode by spatter. The spatter creates a metal film on the glass sealed tube and also causes metal particles to accumulate inside the tube.

Xenon flash tubes are, for example, GP
It has also been used as a pulsed light source for spectrometers according to techniques such as those taught in U.S. Pat. No. 3,458,261 by Bentley et al. The use of high-intensity, short-duration pulsed radiation in this field can obtain a high signal-to-noise ratio when measuring dark and obscure objects, while also heating the object during measurement and causing measurement errors. It has the benefit of being risk-free.

However, using a color spectrometer,
If the tube deteriorates, the life of the flash tube will be significantly shortened, and the spectral characteristics of the light will become unreliable, and the spectral stability will deteriorate. In other words, in conventional xenon flash tube spectrometers, the lifespan of the xenon tube is limited to approximately 100,000 flashes. If it is used more than this, the spectral distribution of the flash light will become irregular, making it impossible to obtain reliable spectroscopic analysis data. While this lifespan is acceptable for systems that operate periodically, it is approximately 10 days for systems that operate continuously, which is far too short for most demands. Become.

SUMMARY OF THE INVENTION According to the present invention, less current is supplied to the xenon tube than in the prior art, while maintaining approximately the same amount of energy per pulse applied to the electrode tube. It has been discovered that such a reduction in current reduces the spatter effect and, as an unexpected effect, produces an electrode that would otherwise degrade in conventional systems.

The new electrodes are formed to provide extremely sharp points. That is, when an arc occurs between this new pair of electrodes, it follows a clear point-to-point path. These electrode points wear out with use and gradually become rounded. As this curvature progresses, the arc path becomes irregular in conventional systems, resulting in irregular light emission conditions that determine the life of the flash tube.

When operated in accordance with the invention, an arc field is created on the electrode surface in which metal nodules accumulate and form new points on the electrode. If continuous flash light emission occurs,
The arc occurs between the same points, either a pair of points originally formed or two new points of metal accumulated during operation of the tube. Regarding arc electrodes, there are papers on various surface effects that have been discovered in the past, but the formation of spheroidal nodules, that is, arc action on the electrode surface, as discovered in the implementation of the present invention, and new points There seems to be no paper that describes its formation.

Another unexpected phenomenon concerns the spectral distribution of flash. That is, when operated in accordance with the present invention, the spectral distribution of the flashes exhibits relatively large differences between successive flash firings compared to the extent observed in conventional systems. Typically, such spectral distribution variations are undesirable and therefore result in flashes that cannot be utilized in spectroscopic measurements. However, when the measured values are normalized at many spectral points (normalization at a large number of points rather than at a single intensity normalization point), the system is It was found that there was little spectral variation. That is, when operated according to the present invention, although the actual spectral fluctuations are large, the fluctuations after multipoint standardization processing are effectively reduced, thereby providing a reliable system.

It has been found that the present invention greatly improves the operating life of xenon flash tubes in spectrometers. In the laboratory, the system suffers from noticeable electrode deterioration and glass tube fogging due to metal film adhesion.
Furthermore, it was possible to continuously perform several million flashes without causing any deterioration in spectral stability.

[Brief explanation of drawings]

1 is a schematic diagram showing a circuit for exciting a xenon flash tube in accordance with the present invention; FIG. 2 is a graph containing a set of curves showing different flash tube excitation characteristics; FIGS. 3A, 3B and 3C are 4 is a schematic diagram showing the spectrometer; FIG. 5A is a graph showing the spectral distribution of a xenon flash tube; FIG. 5B is a graph showing the effect of single point intensity standardization; FIG. 5C is a graph showing the effect of two-point standardization.

DETAILED DESCRIPTION FIG. 1 is a schematic diagram showing the circuitry used to excite a xenon flash tube 1 according to the invention. The flash tube consists of a glass sealed tube 2 into which an electrode is inserted and xenon gas is sealed. Electrode 4 acts as an anode and electrode 3 acts as a cathode. The flash tube is also provided with a whisker 6 connected to the anode 4 and extending outwardly and downwardly toward the tube wall. A film 5 is formed on the outer surface of the sealed tube. This is whisker 6
Starting from the part opposite the free end of the tube, it extends around the sealed tube and descends to the side connected to the cathode. The whiskers and conductive films, which are more fully described in US Pat. No. 3,758,819 to Goldberg, are used to ionize the gaseous medium within the flash tube to generate an arc.

An example of a xenon flash tube suitable for the purpose of spectroscopic analysis is Type 2CP-n manufactured and sold by Scientific Instruments of the United States.

The xenon flash tube is excited by a pulse discharge from the capacitor 10, and one plate of the capacitor is connected to the anode 4 through a series combination of an inductance coil 12 and a diode array 13. The other plate of the capacitor is connected to the cathode 3. The anode of the diode 14 is connected to the cathode 3 of the flash tube, and the cathode of the diode 14 is connected to the anode of the flash tube via the coil 12 and the diode array 13.

The charging circuit for capacitor 10 is transformer 7
and a full wave bridge rectifier 8. The output of this bridge is connected to both ends of a capacitor 10 via a current limiting resistor 11 (100Ω).

The parameters of the circuit are selected to provide a high current pulse discharge to excite the xenon flash tube, thereby producing flashes of reasonable intensity for reflectance or transmittance spectroscopic measurements. The capacitor 10 preferably has a capacitance of 100 μF and is adapted to be charged to a potential of approximately 570V. When fully charged, the capacitor will have about 15 joules of energy with a peak power of about 100,000 watts.

A trigger circuit for initiating arc discharge includes a step-up transformer 22. The high voltage secondary winding of this transformer is connected to the anode 4 and cathode 3 of the flash tube. Further, one end of the primary winding of the step-up transformer is connected to the negative terminal of the bridge 8, and the other end is connected to the positive bridge terminal via a capacitor 21 and a resistor 20. A switch 23 (preferably a solid state switch such as a silicon controlled rectifier) is connected across the series circuit of capacitor 21 and the primary winding of step-up transformer 22. Another suitable trigger circuit is described in Werd US Pat. No. 3,355,625.

Assuming both capacitors 10 and 21 are charged, switch 23 is closed to provide a high intensity short time flash. That is, when the switch is closed, the capacitor 21 discharges and energizes the primary winding of the transformer 22, and the secondary winding is accordingly energized.
A potential that increases toward a high voltage of about 20KV is generated. Diode 13 prevents this high voltage from returning and further discharging capacitor 10. When the voltage applied to the anode-cathode circuit of the flash tube reaches 5-6KV,
The whiskers 6 and the film 5 ionize the gas medium and cause a so-called dielectric breakdown. This generates an arc between the electrodes 3 and 4, which causes the main capacitor 10 to be discharged via the inductance 12, the diode 13, the anode 4, and the cathode 3.

If the coil 12 were omitted, as in the conventional circuit, when the switch 23 was closed, this would result in the discharge characteristic shown by curve A in FIG. This current through the arc electrodes 3-4 is approximately
It rises rapidly to about 5000A and then decreases exponentially to zero. The duration of the main part of this pulse is approximately 20-30 microseconds. This pulse essentially reaches zero in 50 microseconds.

The addition of inductance coil 12 and diode 14 as in the present invention converts the discharge polarity to the form shown in curve B of FIG. 2, thereby having a lower peak current and longer duration. By placing a coil in the circuit,
When the capacitor 10 begins to discharge, energy is first absorbed in this coil. The energy in the coil is then dissipated via diode 14, which is a bypass path for capacitor 10, into the flash tube.

For the particular xenon tube and other circuit parameters described above, the preferred inductance coil is a 3/8 inch diameter core made of 14 gauge wire with a rigid 40
It is wrapped around. The number of coil windings that work satisfactorily is 10 to 100. The corresponding inductance is in the range of 1 μH to 10 μH.
The preferably employed 40-turn coil has an inductance of approximately 3 μH. For larger coils, it is desirable to use layer windings to reduce the resistance of the main current path and reduce the coil size. In a layer-wound coil, the same inductance can be achieved with fewer turns.

When the above-mentioned 40-turn inductance coil is introduced into the circuit, the peak discharge current increases from 5000A to the second
It decreases to about 2000A as shown by curve B in the figure. This current reaches essentially zero in 80 μs. In the case of a spectrometer, it has been confirmed that desirable results can be obtained in accordance with the present invention if the peak current is in the range of 4000-1000A. These pulses have a duration of 60 to 200 microseconds (based on virtually zero value).

FIG. 3A shows a pair of new electrodes 30, 31 used as anode and cathode, respectively, in xenon tube 1. These electrodes are mainly made of sintered tungsten containing impurities such as barium compounds. As can be seen in the figure, these electrodes have sharp points 32 and 3.
2'. When an arc is generated in the flash tube, it flies from one electrode point to the other electrode point. That is, with the new electrodes having relatively sharp points, the flash tube produces a relatively stable arc having a constant length and lateral position.

Figure 3B shows an inductance coil 12 in the circuit.
The electrode of FIG. 3A is shown in profile after operating at a peak current of about 5000 A in a conventional system without using a 3000A peak current. As is clear in the figure, the electrode 33
and 34 deteriorate extremely, and after about 100,000 flashes, the first sharp points appear at the tips of 35 and 36 in the figure.
It slows down like that. This blunting of the electrode causes arcs to emanate irregularly from various locations on the electrode, resulting in variations in arc length and lateral placement, as well as variations in the spectral distribution of flashes. . Such blunted electrodes produce irregular illumination and therefore make spectroscopic measurements unsatisfactory.

On the other hand, when the present invention is actually operated, the electrode regeneration phenomenon discovered as an unexpected effect causes a change in the electrode surface structure as shown in FIG. 3C. During operation of the system, deletion positions occur on the conical end faces of electrodes 37 and 38. The particles of metal material accumulate at this location and gradually form spheroidal nodules as shown at 39 and 40 in the figure. The initial tip point of the electrode is worn and blunted while one of the nodule points, such as 39 or 40 in the figure, is formed as the electrode point for arcing. When one nodule point becomes an arc generation point, nodules in other areas will adsorb metal vapor and grow, eventually forming an electrode point. In this way continuous regeneration of the electrode is possible providing a stable position for arcing in continuous flashing. Since the arc is generated from a specific control point during continuous flashing, the length and lateral position of the arc can be stabilized.

FIG. 4 shows a spectrometer used to perform diffuse reflectance measurements in the visible spectrum to determine the color of a sample. This spectrometer uses a xenon flash tube 43 according to the present invention.
Contains. In order to form a diffuse illumination, the sample 41 is placed in a hollow spherical integrating sphere 42;
Its inner surface is covered with a white diffusion film such as barium sulfate. Irradiation is provided by a pulsed xenon flash tube. This xenon flash tube delivers a short, intense pulse of radiation that drops to zero in essentially 80 microseconds. Because the sample is irradiated for a short period of time, it is possible to measure a moving sample, typically allowing the distance traveled to be ignored during measurements (see, eg, US Pat. No. 3,458,261 to GP Bentley et al.). Also, the use of high-intensity, short-duration pulsed irradiation allows for high-pass filter-type electronic circuitry that is insensitive to ambient light drop-off.

In FIG. 4, the illuminating rays emitted by the light source as rays A, B, and C impinge on the diffusely reflecting wall of sphere 42 and are diffusely reflected therefrom, as shown for example for ray B. A portion of these diffusely reflected light illuminates the sample, while most of the remaining light is incident elsewhere within the sphere. This process is repeated until all the rays are absorbed by the sample or sphere wall. That is, the light reflected by the sample is emitted outside the sphere through the circular aperture 44. The aperture 44 is arranged to pass reflected light rays at a small angle, for example 8°, to the sample normal.

The light beam reflected from the sample is focused by a lens 45 and converged through a slit 46. The purpose of the slit 46 is to limit the spread angle of the rays passing through the optical system thereafter.
The light beam passing through the slit 46 is collimated by a lens 47 and impinges on a dispersive element 48 . A prism or a diffraction grating can be used as this dispersive element. FIG. 4 shows an example in which a reflective diffraction grating is used as a preferred dispersive element.

The diffraction grating 48 performs spectroscopy by dispersing the incident light beam at a specific angle for each of its component wavelengths. For example, red light having a wavelength of 700 nm is emitted as light rays R and R' in the figure, and violet light having a wavelength of 400 nm is emitted as light rays V and V' in the figure. Lens 49 focuses these rays onto a linear array of discontinuously arranged photodetectors 50, focusing the red rays at point R'' and the violet rays at point V. wavelength is a point
It is converged between R″ and V″. the result,
A visible spectrum image is formed in the plane of photodetector array 50 .

The combination of lens 47, grating 48 and lens 49 can replace it with a single grating formed on a concave surface. This concave spherical surface acts as a mirror that can converge light rays. The use of such a concave grating is therefore completely equivalent to the two lens and diffraction grating combination shown in FIG. Also, one of the two lenses can be replaced with a concave mirror. This is because the concave mirror performs the same imaging function as the convex lens it is replacing.

A silicon photodiode can be used as a photodetector. Each photodiode measures only a narrow band. This bandwidth is a function of the width of slit 46 and the width of each photodiode. The measured wavelength will be a function of the photodiode's position in the array. The number of photodiodes arranged in an array is equal to the number of wavelengths to be measured simultaneously. In a typical configuration, e.g.
CIE standard (French International Commission
on illumination) between 400 and 700 nm.
Sixteen detectors are used, evenly spaced at 20 nm intervals. The width and center-to-center spacing of each detector is known to affect measurement accuracy for some colors. Therefore the ratio of detector width to detector center distance is 0.6
The most favorable results can be obtained with a range of ~0.9, and preferably around 0.8.

A pair of reference photodetectors 6 are installed in the hole provided in the spherical wall.
2 and 63 are arranged to monitor the intensity of the irradiation pulse. 5th A~5
As will be explained in more detail below with respect to Figure C, these detectors monitor the intensity of different wavelengths and are therefore equipped with appropriate optical filters. The signals derived from these detectors are used to standardize the signals derived from detector array 50.

A portion of the spherical wall known as the mirror port 53 is
It can be removed by a hinge mechanism 54 to prevent the sample from being illuminated by light beams reflected from the mirror surface (ie, the portion acts as a mirror) and from being measured. When specular port 53 is removed, optical trap 55 prevents light exiting the hole in the bulb wall from being deflected near the outside of the bulb. The center of the mirror port is located 8° from the sample normal, which is symmetrical to the center of the aperture 44 with respect to the sample normal.

A prism 56 is inserted in the path of the sample reflected beam to maintain accurate calibration of the spectroscopic analysis. This prism redirects the light rays from the sample out of the plane of the page of FIG. 4 and thus misses the converging lens 45. However, on the contrary, the light beam reflected from the portion of the sphere wall above the sample is guided to the converging lens 45 and analyzed. Since the reflectance of the sphere wall surface does not vary from day to day and is stable, this measurement can be used as a means of periodic calibration.

FIG. 5A shows the spectral distribution of light from a xenon flash tube illumination source with tungsten electrodes. As is clear from this figure, the intensity varies depending on the wavelength. The light appears to contain both certain narrowband and broadband distributions.

For simplicity, only eight detector measurements a-h are shown in FIG. 5A, although as previously noted, a typical system will make 16 or more measurements.

One of the detectors 62 (Figure 4) is positioned to monitor the light intensity at the wavelength shown at 70 in Figure 5B. This measurement is used by signal processing electronics 51 (FIG. 4) to normalize intensity variations. This is obtained by dividing the measurement obtained from detector 50 by the reference measurement obtained from detector 62.

As a result of observing the operation of flash tubes, it was discovered that spectral fluctuations are accompanied not only by fluctuations in intensity, but also by a spectral perturbation effect in which the light intensity at one end of the spectrum is sometimes greater than the light intensity at the other end. It was done. Thus, for example, if a single point intensity normalization is achieved at wavelength 70, a deviation from the true value will occur at other points in the spectrum. Generally 5B
As shown in the figure, these deviations have values between lines 71 and 72, as indicated by the shaded area, and increase with increasing distance from the monitoring point.

To compensate for spectral shifts in light intensity variations, it is desirable to further standardize at at least one additional point, as shown at wavelength 73 in FIG. 5C. A reference detector 63 is used for this purpose. Two monitoring points 70 and 73
should be well separated as shown in the figure.

A preferred procedure for acquiring data for spectral standardization uses standard white tiles and records values for each of detectors 50 and 63 after intensity standardization (by detector 62). With this data,
For each of the different measurements that may be obtained from the detectors 63, the average value (of the plurality of measurements) at each of the detectors 50 is determined. be able to. These average values are arranged in a lookup table and used for spectral standardization. When used with unknown samples, once a value is measured by detector 63, a correction factor corresponding to this measured value is retrieved from the lookup table and used to modulate the value obtained from detector 50.

After all, the two-point standardization technique of the present invention is based on one point in the spectral range (70 in Fig. 5B),
The light intensity fluctuation within the same range is corrected, and other points (5th C
This is to correct the spectral shift of the light intensity fluctuations corrected based on FIG.

By using the two-point standardization technique described above,
Accurate values are guaranteed at the normalization points of wavelengths 70 and 73, yet the degree of intensity variation is as expected at other points in the spectrum. This variation is shown in enlarged form at wavelengths 74 and 75 in FIG. 5C.

When the flash tube circuit is modified according to the present invention as shown in FIG. 1, a relatively large spectral shift is created. As shown in Figure 5B, the variation in the presence of inductance is at wavelengths 76 and 77.
However, according to the present invention, it is expanded to between lines 71 and 72, where there is a relatively large maximum variation d ₁ , which appears to be a more unstable irradiation spectrum and unfavorable for spectroscopic measurements. It can be considered.

However, even if the variation after intensity standardization is relatively large with the modifications according to the present invention, the variation after two-point standardization is significantly reduced. Fifth
As shown in Figure C, the maximum variation decreases from _d2 to _d3 . It is unclear why this unexpected effect is obtained.

Color spectrometers are routinely evaluated according to their ability to repeat the same measurements. These ratings have a value of 1.0
The values are based on the value of color difference when the color difference is within the limit that can be detected with the naked eye. Repeatability is the RMS over a series of measurements on the same sample
The values are different colors.

Single point intensity normalization by conventional systems results in color difference values as high as 1 or 2. 2
A similar system of the present invention that performs point standardization results in a small difference value range of 0.17 to 0.25. According to the present invention, repeatability after two-point standardization is improved to a color difference range of about 0.09 to 0.15.

Signal processing electronics 51 preferably include a sample and hold circuit and each detector 50, 62 and 63.
and a read-only memory (ROM) for the lookup table, and a microprocessor programmed to perform the standardization calculations described above. The sample and hold circuit is controlled to provide a measurement window corresponding to a light pulse duration of 60 to 200 microseconds, and in the preferred embodiment shown by curve B in FIG. 2, about 80 microseconds.

Optionally, hard-wired digital logic circuits or analog computing systems can also be used.

The preferred embodiments of the present invention have been described above. However, various modifications may be made within the scope of the invention, which is limited only in the scope of the appended claims.

Claims (1)

[Scope of Claims] 1. Irradiating the sample 41 under measurement with a xenon flash tube 43 energized by a current pulse having a peak current of 4000 A or less and a duration of 80 to 200 μs; generating measurements by detecting the light after it has been modulated by the sample 41 at a plurality of different wavelengths;
generating a reference value by detecting light from the flash tube unmodulated by the sample 41 at at least two different wavelengths; and generating a reference value by dividing the measured value by one of the reference values. standardizing the intensity variations; and
modulating said measured value in accordance with a lookup value corresponding to the other of said two reference values, said lookup value being based on data obtained by a reference sample; A method of color spectrometry measurement, characterized in that the value is an average corrected value of intensity normalized measurement values at the plurality of detection wavelengths that have a correlation with the measurement values of the other reference detector. 2. The method of claim 1, wherein the peak current is about 2000 A and the peak width is about 80 microseconds. 3. A capacitor discharge circuit including a xenon flash tube 43 having tungsten electrodes 3 and 4, a capacitor 10 connected to energize the flash tube 43 during its own discharge, and the capacitor 10 and the flash tube 43. After the light from the flash tube 43 and the sample 41 under test are modulated by the inductance 12 for reducing the peak current inserted into the discharge path between the two, the light from the flash tube 43 is arranged to measure different wavelength components. at least two reference detectors 62 for measuring light from the flash tube 43 unmodulated by the sample under test 41 at at least two different wavelengths; 63, and in response to the signals from said detector 50, the values of those signals are transmitted to said two reference detectors 6.
2,63 values, and according to the lookup value corresponding to the other of the two reference detector values 62,63. and an electronic signal processing means 51 for correcting the spectral shift of the variation by modulating the signal from the device 50, the inductance value of the inductance 12 inserted in the discharge path is 1
A color spectrometer characterized in that the peak current of discharge is reduced to within the range of 1000 to 4000 A by selecting the value within the range of ~10 μH. 4. The color spectrometer according to claim 3, wherein the inductance reduces the peak current to 2000A. 5. The color spectrometer further includes a diode in a bypass circuit for the capacitor such that substantially all of the energy from the capacitor is transferred to the flash tube via the inductance. A color spectrometer according to claim 3 or 4. 6. The color spectrometer according to claim 3, wherein the measurement window for the detector is set to approximately correspond to a flash period within a range of 60 to 200 μsec. 7. The color spectrometer according to claim 6, wherein the measurement window is set to about 80 μsec.