Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS2577815 A
Publication typeGrant
Publication dateDec 11, 1951
Filing dateDec 15, 1949
Priority dateFeb 27, 1946
Publication numberUS 2577815 A, US 2577815A, US-A-2577815, US2577815 A, US2577815A
InventorsCaldecourt Victor J, Peterson Eugene W, Saunderson Jason L
Original AssigneeDow Chemical Co
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Photoelectric instrument for direct spectrochemical analysis by the internal standard method
US 2577815 A
Images(5)
Previous page
Next page
Description  (OCR text may contain errors)

Dec.-11, 1951 J. L. SAUNDERSON ETAL 7,8

PHOTOELECTRIC INSTRUMENT FOR DIRECT SPECTROCHEMICAL 7 ANALYSIS BY THE INTERNAL STANDARD METHOD Original Filed Feb. 27, 1946 5 Sheets-Sheet l n W Z M nm m e w r SJ 6 e 5 o -R w d .Rt RE I a P i L m i e a z W J V E -N MN kNmXQQQ W HW\NN I /WN m %N BER 1951 J. L. SAUNDERSON mm. ,3 5

PHOTOELECTRIC INSTRUMENT FOR DIRECT SPECTROCHEMICAL M ANALYSIS BY THE INTERNAL STANDARD METHOD Original Filed Feb'. 27, 1946 5 Sheets-Sheet 2 Sp ec/ra/ /in@ Jason L. 80unans0n my 5 Vic/0r J Ca/a ecourf E aye/7e W. Peferson v.1. 1,, S AUNDERSON ET AL 2,577,815 PHOTOELECTRIC' INSTRUMENT FOR DIRECT SPECTROCHEMICAL Dec. 11, 1951 ANALYSISBY THE INTERNAL STANDARD METHOD Original Filed Feb. 27, 1946 5 Sheets-Sheet 3 X5035 W QBQ Vic/0r J Ca/a ecour/ Eqyene W Pe/erson Dec. 11, 1951 J. L. SAUN'DERSON ETAL 2,577,815

, PHOTOELECTRIC INSTRUMENT FOR DIRECT SPECTROCHEMICAL ANALYSIS BY THE INTERNAL STANDARD METHOD Original Filed Feb. 27, 1946 5 Sheets$heet 4 ower 65 line 64 66 65 7'0 spark source To background ska/fer and/e verJ/ng re/ay:

To mas/er 3/: u//er To .s/ror/ing re/ays To re Cara er mo/or re/oy To measuring re/ay: 7'0 /e0 back re/ag 7'0 recor d'ng ,o ens From power pack From re/rence Jasan L. 60unaers0n V/c/ar-J Ca/oeaour/ E 1.19am? Ml Pe/ergson Dec. 11, 1951 J. 1.. ,Js'AuNDzRsoN ET AL Y 2,577,815 PHOTOELECTRIC INSTRUMENT FOR DIRECT. SPECTROCHEMICAL ANALYSISBY THE INTERNAL STKNDARD METHOD Original Filed Feb. 27, 4946 -5 Sheets-Sheet 5 Q QG $6 $6 m6 N6 QNQWN Wmwm W /6 M yan i SQ w bm m 8 Q saw \v m me kw w m wmwk nmQiNQEQQEQQ \G mmG S V 1 m M Mam m Q9 Patented Dec. 11, 1951 PHOTOELECTRIC INSTRUMENT FOB DI- RECT SPECTROCHEMICAL ANALYSIS BY THE INTERNAL STANDARD METHOD Jason L.' Saunderson, Victor J. Caldecourt, and

Eugene W. Peterson, Midland, Mich., assignors to The Dow .Chemical Company, Midland, Mich., a corporation of Delaware Original application February 27, 1946, Serial No.

Divided and this application December 15, 1949, Serial No. 133,181

4 Claims.

This invention relates to a direct-reading or phototelectric spectrometer for use in spectrochemical analysis. It concerns improvements both in the optical system and in the electrical measuring circuits of such an instrument.

This application is a division of our co-pending application Serial No. 650,676, filed February 27, 1946.

Quantitative spectrochemical analysis by the internal standard method, an exceedingly valuable tool for the metallurgical industries, has heretofore been carried out almost entirely by photographic means, such as those described, for example, in U. S. Patents 1,979,964 and 2,043,053. This method, while of broad usefulness, is, for routine process control, disadvantageous in that a considerable period of time, at least several minutes and often much longer, elapses between the exposure of the photographic plate and its development and interpretation.

It has been appreciated that a considerably greater rapidity of analysis might be realized if the spectral lines to be compared, instead of being recorded photographically, were focused on separate photocells, the relative outputs of which could then be measured electrically. However, the photoelectric spectrometers which have been proposed leave much to be desired in the way of simplicity of operation, as well as in mattersof inherent accuracy and of stability in the presence of vibration and ordinary environmental changes.

-It is therefore an object of the invention to provide an improved direct-reading spectrometer which may be made fully automatic and which is usually capable of giving complete analyses for as many elements as desired in a period of less than one minute. Another object is to provide an instrument of this character which may be operated continuously over long periods by nontechnical personnel and yet exhibit an inherent accurancy equal to or exceeding that obtained with the spectrograph. A further object is to provide a spectrometer in which the relative accuracy of analysis for any given element is essentially constant over wide ranges of concentration of that element.

Another object of the invention is to provide an instrument embodying improved electrical means for integrating and comparing the relative outputs of two or more photocells when each is exposed to light of continuously varying intensity. Additional objects are to provide optical means for compensating for background radiation which may be received by a photocell together with a desired spectral line and electrical means for balancing out the dark current of the cell. A further obiect'is to provide an improved circuit for determining'the electric charge imposed by a photocell on a condenser, such circuit including a highly stable direct-current electrontube amplifier.

Still other objects and advantages of the invention will be apparent from the description to follow, which is made with reference to the accompanying drawings, in which Fig. 1 is a simplified diagram illustrating the inter-relation of the essential elements of the apparatus;

Fig. 2 is a horizontal section through the photocell compartment of the spectrometer, showing the arrangement of the optical elements;

Fig. 3 is a sketch illustrating the centering of a spectral line in an exit slit of the spectrometer, the line being shown enlarged with respect to the width of the slit;

Fig. 4 is a similar sketch illustrating the positioning of a spectral line in an exit slit when a background compensating prism is used;

Fig. 5 is a sketch showing the details of another slit arrangement for compensating for background spectral radiation;

Fig. 6 is a horizontal section through a photocell and its diffuser;

Fig. 7 is a diagram of the electrical measuring and recording circuits of the apparatus;

Fig. 8 is a circuit diagram of the automatic tim ing system of the spectrometer;

Fig. 9 is a detailed circuit diagram of one of analysis of the PRINCIPLE 1 In the apparatus of the invention, light produced by sparking electrodes of the sample to be analyzed is resolved by a spectrometer which focuses characteristics spectral lines of the'reference and unknown elements on sensitive photocells, causing the fiow of minute photocurrents. The varying intensities of the lines during the sparking period are integrated by storing the cor-- respondingly varying photocurrents in condensers. After the sparking period, these stored charges are compared by allowing the condensers to discharge through similar resistances and measuring the difference in the times required for such discharges to take place. Electron-tube amplifiers follow the course of the condenser discharges and acuate relays which operate a recorder, the indications of which can be calibrated directly in terms of the analysis of the sample being sparked.

SIMPLIFIED DIAGRAM The essential elements of the new apparatus may be explained with reference to the simplified diagram of Fig. l.

The material to be analyzed is formed into electrodes H which are caused to spark by connecting them across a high voltage source II. Light from this spark is focused by a lens it through an electrically-opened normally-closed master shutter 14 onto the entrance slit ll of a spectrometer it. The entering light falls on a concave grating H, which forms images of the slit along the Rowland circle, the positions of these spectral lines being dependent on the wavelengths present in the light. Exit slits II are located along the circle in positions to allow passage of selected lines corresponding to the internal reference element and the element being analyzed for in the electrodes I I. These selected lines are focused by lenses is on the sensitive elements of electron multiplier photocells It, the operating potentials of which are supplied by a hi h voltage source 2|.

During sparking of the electrodes, the spectral line falling on each photocell causes a 'corresponding fiow of current, which is stored in a condenser 22 connected in the circuit through a two-way switch 24. Later, the relative magnitudes of the charges thus accumulated are measured by connnecting each condenser across its individual discharging resistor 23 by throwing the two-way switches 24. when these switches are thrown, the falling potentials of the condensers are followed by electron-tube direct-current amplifiers 25 in the output circuits of which are voltage-responsive relays 28 and 21 with their output terminals in series. The relay 2! in the circuit corresponding to the spectral line of the unknown element is of the normally-closed type, while the relay 21 in the circuit of the reference element is normally open. Whenever operation of the amplifiers holds the contacts of both relays closed at the same time, the pen circuit ll of a moving-tape recorder 28 is energized by a source III, causing a mark to appear on the tap In practice, it is necessary to choose the reference spectral line, or to design the circuits, so that during sparking the reference condenser reaches a higher potential than the other condensers. The relays and amplifiers are adjusted so that the relays are actuated whenever the potentials across the condensers exceed a predetermined low value, conveniently about 1.0 volt.

In making an analysis, the switches 14 are set to connect the photocells to their condensers, and the sample to be analyzed is sparked. The resulting spectral lines falling on the photocells a cause photocurrents which charge the condensers 22 at rates proportional to the intensities of the incident lines. After the sparking period, the switches 24 are thrown simultaneously, whereupon the condensers I! begin to discharge through their resistors 23. As soon as the switches close, the amplifiers ll detect condenser voltages well over the predetermined low value,

and actuate both relays 20 and 21, leaving the recorder circuit still open. As discharge of the condensers continues, the falling potential in the condenser corresponding to the unknown element reaches the predetermined low value first, whereupon the relay II is de-actuated and closes. Current at once fiows in the recorder circuit II. making a visual record on the moving tape. When, after an interval of time, the falling potential across the reference condenser also reaches the predetermined value. the relay 21 is deactuated, and the recorder circuit is again broken. As a result of this sequence, the length of the mark on the moving recorder tape is a direct indication of the difference in times of discharge of the reference and the unknown condensers. This difference is a function of the concentration of the unknown element in the sample, as will be evident from the following considerations.

GENERAL CONSIDERATIONS Theory where E0 is the initial voltage. E is the voltage after time t, and e is the base of natural logarithms. Considering two condensers of capacities Cu and Cr, having initial voltages Eu and E: and discharging through resistances Ru and Br. the times in and tr required for the condensers to reach a predetermined low voltage E. are. from Equation 1:

(a) ta=Raca 108s Ea/Es (3) tr=Rrcr 108s lr/E'a If the resistances and capacitances are chosen so that their products are equal to some selected value RC. that is. if

(4) RaCa=R1Cr=RC then, by combining Equations 2, 3. and 4, and taking tr-ta=At, it follows that (5) At=RC 108s Er/Ea But the initial voltages Er and Eu on the condensers are proportional to the total charges, and hence also proportional to the integrated intensities Ir and In of the spectral lines. whence. sub

stituting in Equation 5, there appears:

'(6) At=RC log bla-i-constant From spectrochemicai theory, it is known that the ratio of the intensities of the spectral lines of the reference and unknown elements is approximately proportional to the ratio of the concentrations x. and x of these elements. Since the reference element is essentially of constant concentration, it follows from Equation 8 that Aim-RC log. xe+constant gamers From these considerations, then, it is seen that a plot oi At, the difference in times or discharge of the reference and unknown condensers, against the logarithm of Xu, the concentration of the unknown element, is essentially linear.

In the apparatus of Fig. 1, this time diflerence. At, is, as explained, read directly from the length of the mark on the recorder tape. In making analyses, then, the apparatus is first calibrated by determining the length of the recorder mark for several samples in which the element being analyzed for is of known concentration. Then a line drawn through a plot of these lengths against the logarithms of the concentrations of the element gives the desired calibration chart. From this, analyses of unknown samples can be run by simply sparking the sample according to Several important advantages arise from the fact that, in the apparatus of the invention, the quantity actually measured, viz. the difiference in discharge times of two condensers, varies logarithmically with the concentrations to be determined. In the first place, the spectrochemical method in general tends to have an essentially constant relative accuracy for the quantitative determination of an element over wide ranges of concentration. For instance, the absolute error in determining an element'at per cent concentration is roughly about ten times as great as the absolute error at 1.0 per cent concentration. The logarithmic calibration curves of the present apparatus are similar in this respect, in that they may be read with constant relative accuracy over the whole scale. A second advantage, apparent from Equation 7 above, is that the slope of the logarithmic calibration curve is independent of changes in apparatus conditions, since all proportionality factors fall into the constant in the equation. Thus, a change in the sensitivity of a photocell circuit, or dirt in the optical path, will only shift the calibration curve parallel to itself without altering its slope. In consequence, if a change in the apparatus occurs, it is necessary, in order to recalibrate, to run only a single standard sample, to give a At value which, with the constant'known slope, will give the complete new calibration curve.

An additional advantage of the apparatus of the invention is that, although the instantaneous intensities of the spectral lines, and hence also the instantaneous photocurrents, may vary considerably due to flickering of the spark, the condensers, in storing these currents, effect a complete integration over the time of sparking. As a result, when the charges on the condensers are measured, there is obtained an integrated relative intensity value for the two spectral lines which is practically independent of variations in both the total intensity of the spark and the time of sparking. Since this integration is quite similar to that efiected by the photographic plate of a spectrograph, the method of the invention permits use of virtually all the established techniques or the internal standard method so well known in spectrographic analysis.

Background compensation there is, at every wavelength, a certain amount of background radiation. In the apparatus of the invention, some part of this background enters the photocell together with the line on which it is focused. As long as the intensity of the line is high relative to that or the background, the latter is of little consequence. However, when the cell is focused on a line of weak intensity, as, for instance, when analyzing for an element present in low concentration, the background may bea substantial part of the total light incident on the cell. While this circumstance does not render analysis impossible, since the apparatus is calibrated empirically, the relative accuracy of analysis will be reduced, largely because the intensity of the background tends to vary somewhat from sample to sample.

In the invention, this effect of background radiation'may be largely or wholly overcome by an optical system including a shutter which permits repeatedly exposing the photocell first to the desired line together with its background and then to a selected portion of the background alone. At the same time, operation of the electrical circuits is synchronized with the shutter so that the condenser associated with a photocell, while being charged by the cell each time the latter is exposed to the line and background, is also allowed to discharge through the cell each time the latter is exposed to the background only. By making the times of charging and discharging equal, the additive effect of the photocurrent due to the background during charging is nullified by its subtractive efiect during discharging. At the end of the sparking period, then, the condenser charge contains little or no component attributable to the background, and is a function only of the intensity of the desired line.

Dark current compensation The method of background compensation just discussed has the further effect of also compen-' eating for the dark current of the photocells, thereby rendering frequent adjustment of the cell circuits unnecessary. However, an alternative method of dark current compensation may also be desirable, especially for use in those instances where background compensation is not required. Such dark current compensation may be accomplished by connecting a high resistance across the photocell at the time it is charging its condenser, and arranging this circuit so that the current flowing through the resistance exactly equals the dark current. Balancing means for making this adjustment are provided and will be described later.

Direct current amplifier In the present spectrometer, the direct-current amplifiers which follow the discharge of the photocell condensers must have a high gain in order properly to actuate the relays in their output circuits. On the other hand. they must also 7 be accurate and stable, preferably with so little drift that they require adjustment only once or twiceaday. Thisisachieved.accordingtothe invention, by means of two-stage direct-current amplifiers which are provided with feed-back networks to render them highly degenerative and hence stable. .The amplifiers are normally maintained in the degenerative state. However, at the moment of discharging the photocell condensers. the feed-back networks are temporarily broken, converting the amplifiers to a sensitive orhigh gain condition. After the discharge, the feedback circuits are again closed, returning the ampliher to a stable state in which any tendency toward drift is compensated.

DESCRIPTION OF APPARATUS are not shown. Adequate descriptions of these elements, as well as methods of preparing the electrodes, are given in J. Opt. Soc. Amer. 34, 104-5 and 116-9 (1944).

Optical details ,The spectrometer housing II and the grating 11 (Fig. 1) are of known type. It is important that the spectrometer be rigid and solidly mounted, and desirable that the housing be maintained at a constant internal temperature, conveniently 90 F., as by a thermostated circulating air heater.

As shown particularly in Fig. 2, the spectrometer exit slits it are formed in thin metal plates II which are adjustably mounted over slots 32 in the spectrometer housing It along its focal curve II in positions to pass the desired spectral lines. In general, the width of an exit slit should be about two times that of the entrance slit for those lines, such as the reference line, on which no background correction is necessary. and about four times on lines on which a correction is to be made.

. If no background correction is needed, as in the case of the line of the reference element, the line is approximately centered in the exit slit ll (Fig. 3). However, when a background correction is made, as in the case of the line of an unknown element, the line is directed through one side of the slit, allowing the background radiation to pass through the other side (Fig. 4). Inthis case, a small prism 3| is centered just behind the slit so as to split the light into two beams, one of which includes the desired line and the other contains only the background. Each of these beams enters the lens I9 and is focused on the photocell 20 (Fig. 2). The purpose of the prism is to separate the two beams sufficiently to allow satisfactory operation of the background compensation mechanism.

This compensation is effected by means of a shutter interposed between the exit slits and the lenses and consisting of a curved metal strip having slits 36 cut in it in positions corresponding to the exit slits and being somewhat wider than the latter. This shutter is mounted to reciprocate in a fixed guide 31, and is moved 8 backandforthbyopposedsolenoidsflwhich are attached to either end. Movement of the shutter is limited by stops it one of which is adjusted so that at the limit of shutter travel in one direction the shutter is in a position to allow free passage of the spectral lines from the exit slits to the lenses and to block the background beam (as shown in Fig. 2). The other limit of travel is such that the shutter is in a position in which, in the case of lines requiring no background compensation, it cuts oil all light to the photocell, and in the case of lines requiring correction, it cuts off the spectral line but allows free passage of the background beam. The shutter is reciprocated by alternately activating the solenoids in synchronism with the photocell condenser reversing circuits, as will be later described.

The exit slits, prisms, shutter mechanism, lenses, and photocells are all mounted rigi ly on the frame of the spectrometer by means not shown, and are enclosed in a light-tight case I. The lenses and prisms are of quartz or other material transparent to ultraviolet light.

An alternative method of background correction, useful when the immediate background of a selected line contains interfering lines, is shown in Fig. 5. In this case, a second exit slit 4| is located near the slit is for the desired line, but in a non-interfering spectral region, and the shutter 35 is adjusted to cut off beams from the two slits alternately. Both beams are, of course, focused on the same photocell. 1

While Figs. 1 and 2 illustrate a spectrometer with onh! two exit slits and photocells, one for the reference element and one for the unknown, it will be appreciated that any desired number of exit slits and photocells may be arranged along the focal curve at positions corresponding to various spectral lines to permit determination of several unknown elements at once, using the same; or more than one, reference spectral line. In such cases, certain of the beams passing through closely adjacent slits may, if desired, be deflected by mirrors before entering their corresponding lenses and photocells, thus avoiding crowding together photocells directly behind the focal curve.

The considerations involved in selecting appropriate spectral lines for any given analysis are similar to those in spectrographic work, except that greater latitude in choosing lines is possible. (See discussion of Saunderson et al. in J. Opt. Soc. Amer. 35, 681 (1945). a paper which also describes the present invention.)

Photocells While any of a number of types of photocells may be employed in the apparatus, optimum results are obtained with phote-emissive tubes, particularly electron multiplier photetubes, such as the 931A, which combine extreme sensitivity with high amplification. These latter phototubes, however, present'a special problem because of the great variationin output occasioned by slight changes in the position of incidence of the spectral line.

Thus, as will be appreciated, from Fig. 1, while the positions of the spectral lines of the elements are fixed, the light beams, after passing the exitslits, move with every motion or fiicker of the spark between the electrodes Ii. The lenses it, in focusing the beams on the photocells III, eliminate the eflfects of motion in the light source except for aberrations resulting from using difquartz plate 42 mounted a short distance in front of the envelope 43 oi the photocell, the space between the plate and the envelope being surrounded by a light shield 44. With this construction, critical positioning 9! the phototubes is not required.

Measuring circuit I The basic electrical measuring circuits of the apparatus may be explained with reference to Fig. 7. For simplicity, the diagram shows only three multiplier phototubes and measuring circuits, for determining the relative intensities of two unknown spectral lines relative to the same internal standard line. In Fig. 7, the timing circuits for actuating the various relays are not shown, and the electron-tube amplifiers are indicated in simplified form. The circuit is shown in its normal position before an analysis is run.

The operating potential for all the electronmultiplier tubes is supplied by a 1000-volt regulated power supply as constructed essentially according to the Miller design (Electronics, November, 1041, p. 29). The 250-volt amplifier power pack 56 is of a similar type.

The dark-current of each phctotube 20 is compensated by a leak circuit in. which the photocell anode ill ,is connected through a high re-- sistance as (1000 megohms) to a potentiometer it (100 volts), one terminal of which is grounded.

During the sample sparking period, the output of each photocell, except for the small darkcurrent compensation. passes in series through a manual test switch 50 and the normally-closed contacts of a relay to its storage condenser 22, the return circuit being through the normally closed contact of a relay 54. Actuation of the relay 54 interposes a standard cell 52 in the return circuit. The polarity of the condenser may be changed by a reversing relay 52, and the condenser is shorted, except during measurements, by another relay 53.

At the end of the sparking period, each condenser may be isolated irom its photocell and connected directly across its discharging resistor 23 by operation of the relay 5|. As this operates, each condenser is connected through a bias cell 55a to the input grid 55 of the corresponding amplifier 25, which in turn actuates a corresponding recorder relay 26 or 21.

Each amplifier 25, which will be more fully described later, comprises two stages and is stabilized by a feedback circuit 56 which may be broken by a relay 51. The magnitude of the amplifier output is indicated visually by an electron-ray tube 58, the ray-control grid of which is connected in the plate circuit of the'second amplifier stage.

The recorder relays 28 and 21 in the amplifier outputs control the recorder pen circuits 28, as

already explained with reference toFig. l. The

recorder, which is conveniently oi the movingtape type described in Patent 2,251,742, is driven by a self-starting synchronous motor 59, the

they have been discharged rapidly, condensers 10 using polystyrene as the dielectric being especially suitable (J. Opt. Soc. Amer. 35, 690 (1945). The electrical magnitudes of thecapacitance of each condenser (0.1 to 1.0 microtarad) and of the resistance 01' each discharging resistor (5 to 50 megohms) are chosen so that their mathematical product is of the order of several seconds and is substantially equal for all condenserresistor pairs in the apparatus, in order to satisfy the mathematical requirements 01 the system, as in Equations 4 and 5 above.

Operation of circuit As long as the relays are in the positions shown in Fig. 7, the input grids 55 of the amplifiers 25 are all maintained at a fixed potential below ground potential determined by the bias cells 55a. With this condition obtaining, the operator first adjusts each amplifier-relay circuit so that each indicator tube 58 or magic eye is just "closed," and so that each recorder relay 26 or 21, though still in the normal position, will be activated whenever the corresponding amplifier input-grid potential is depressed below the aforesaid bias potential.

' With the amplifiers thus normalized, and with the photocells 20 receiving no light, dark current compensation is made. Each test switch 50 is thrown, connecting each photocell directly to its amplifier input grid. At the same time, the feedback relay switches 51 are-opened to place the amplifiers in the high-gain state. ,Each voltage-divider 49 is then adjusted 'until the corresponding indicator tube 58 is again closed." The entire dark current of each photocell is thus balanced out, and fiows through its leak resistance 188. The switches 50 and 5'! are returned to normal positions, and the measuring circuits are ready for use.

In making an analysis with the spectrometer utilizing the measuring circuit of Fig. 7, the sample to be analyzed is first sparked for a few seconds, after which the exposure period is initiated by a timing system later to be described which actuates the master shutter I4 (Fig. 1) and the condenser shorting relays 53 (Fig. '7) at the same instant. The light 01' the spark then enters the spectrometer and the resulting spectral lines fall on the photocells 20, causing photocurrents to flow; the condensers 22, simultaneously unshorted, store these photocurrents throughout the sparking period. As sparking proceeds, the solenoids 38 for reciprocating the background compensation shutter (Fig. 2) and the condenser reversing relays 52 (Fig. 7) are operated in synchronism. Thus, each condenser is alternately charged with photocurrent corresponding to the intensity of the spectral line plus background and discharged for an equal period with photocurrent corresponding to the intensity of the background alone.

When the exposure period has ended, the timing system closes the master shutter ll, cutting oil. all light to the photocells. It simultaneously places the background shutter solenoids 38 and the condenser reversing relays 52 in normal positions, and closes the recorder motor relay 60, setting the recorder roll Si in motion. The condenser measuring relays 5| and 5| and the feedback circuit relay 5! are then actuated simultaneously, connecting the condensers 22 across their discharging resistors 23, isolating the amplifier input grids 55 from ground, connecting the standard cell 62 in circuit, and converting the amplifiers to a high-gain condition. Assuming that the potentials on the charged condensers all 11 exceed in magnitude that of the standard cell. the amplifier input grids it instantly detect potentials below the bias potential, and the amplifiers actuate the recorder relays and 21. The

recorder pen circuits is for the unknown elements remain open, but the circuit for the reference element, involving the relay 21, is at once closed, so that recording begins on this one pen. Then, when the decreasing potentials on the condensers of the unknown elements reach voltages equal in magnitude to that of the standard cell. i. e. when the grids ll reach bias potential,

the relays It are released, and the remaining recorder pen circuits close and cause marking of the recorder tape. As soon as the decreasing potential on the reference element condenser also M l s that of the standard cell, its amplifier deactuates the relay 21, and the pen circuits are again broken, so that all recording stops. The condenser discharge relays II and ll, the condenser shorting relay the amplifier feedback relay I1, and the motor relay it are then all released, placing the circuits in the normal or inactive condition, as in Fig. 7, and returning the amplifiers to the degenerative or self-stabilizing state. The entire apparatus is then ready for another analysis.

Timing system The details of the timing mechanism for initiating the operating impulses for the various relays in the measuring and spark circuits of the direct-reading spectrometer are shown in Fig. 8. The controlling elements of the mechanism are two timers and a series of cam-operated switches, togetherwith several relays. The cams rotate continuously when the spectrometer is in use, while the other elements function only in a predetermined sequence each time the "start" switch is operated to make an analysis.

In the circuit of Fig. 8, which is shown in its normal or inactive condition, the current is drawn from a line source 0, the return circuit bein in all cases through ground. The current first flows through a normally closed cut-out relay M to the "start" switch ,and from there to the coil of a self-locking normally-open starting relay This starting relay, when closed, supplies current through a line I] to operate the spectrometer spark source l2, and also to actuate a first timer switch 6.. This latter is set to allow the elapse of several seconds for the electrode spark to stabilize itself before starting the condenser charging cycle.

When the timer switch ll closes to start the spark exposure period, potential is supplied from the line It to the start cam switch as, the cam of which is mounted on a shaft It turned continuously by a geared synchronous motor II at a slow rate, conveniently once very two seconds. As soon as this cam switch it closes, the control impulse fiows through a lead 12 to the coil of a multi-contact self-locking starting relay l3.

- When this starting relay is actuated. power flows 1! supply circuit II and the lead It to the relay I. operating the spectrometer recorder motor ll (Fig. 7). The entire system of Fig. 8 then remains in this condition until'the exposure period is ended by operation of the second timer ll.

As soon as this second timer switch I! closes. current fiows' from a power supply lead I2- lIa-lib to energize a quick-closing delayedrelease control relay 88, which in turn actuates the coil of a multi-contact stopping relay ll. When this latter relay II is energized, it connects the power supply lead ll-OIc-"c to the coil of the control relay 0, thus locking both the control and the stopping relays in the actuated position. This same actuation of the stopping relay it also connects the power line I to the power supply circuit ll for the recorder motor. which circuit is connected by a Jumper it to the lead II for the condenser shorting relays. thus preventing the latter from being deactuated. Likewise, actuation of the stopping relay ll closes a circuit ll connecting the stop cam switch IT with the ground side of the coil of the starter relay II. Thus, as soon as the stop cam rotates to close the switch 01, the latter shorts out the coil of the starter relay II, deactuating the latter. The starter relay II at once drops out, breaking the power supply lead 14 to the master shutter It, thus terminating the exposure period, and also allowing the timer It to reset.

This dropping out also connects the power supply circuit 00 to the lead 0| to the recorder motor relay (Fig. 7),, starting the motor 59.

This same current supplied to the recorder lead It also flows through a lead II to the coil of the cut-out relay l4, actuating it and breaking the power supply to the locked starting relay N. This latter then opens, shutting oil the power supply to the spark source I! and the first timer switch ll, allowing the latter to reset.

The current in the recorder motor lead It is also transferred by the lead 88 and a second lead I! to the measuring cam switch so, which, when it rotates into closed position, energizes the coil of a multi-contact self-locking measuring relay 9|. This latter, on closing, connects power from the source lead I! to leads 92 for the condenser measuring relays II and II, and the amplifier feedback relays I! (Fig. 7), thus starting the measuring cycle of the instrument. This closing of the measuring relay II also breaks the connection between the power lead I! and the lead 82a, but it is simultaneously re-established by the closing of auxiliary contacts on the reference amplifier recording relay 21, which is actuated by its amplifier at the same instant that the measuring cycle begins, as already explained with reference to Fig. 7.

When the measuring cycle is completed, the reference amplifier releases its relay 21, thus breaking the power supply to the power lead Ila-82c and de-energizing the coil of the control relay II. This delayed-release-relay is set to remain closed a few seconds to allow the recorder motor time to run out extra tape, after which the relay opens. When it opens, the stopping relay I4 is thus deactuated, and shuts of! the power to the recorder motor leads I0 and II. stopping the motor and de-energizing both the cut-off relay N and the measuring relay OI, returning all the circuits to the normal state ready 7 for the next analysis.

As shown in Fig. 8, the cams of the starting switch it and the stopping switch I! provide only momentary closing of their respective circuits, and are set so that they'operate simultaneously. The reversing cam switches 16 close their circuits alternately, each circuit remaining closed for one half revolution of the camshaft. The particular reversing cam which holds the background shutter in position to transmit the spectral lines is timed to close simultaneously with the start and the stop cams. The cam of the measuring switch 90, which likewise closes its circuit only momentarily, is set to lag the stop cam by an appreciable interval, such as onefourth of a revolution, to allow the recorder motor time to reach speed before the measuring cycle begins.

The circuits of Fig. 8 render operation of the spectrometer entirely automatic. .Thus, the operator, after inserting the electrodes to be analyzed into their holders, merely pushes the start button 65. The spark' across the electrodes is at once turned on. Then, after a warmup interval determined by the timer 68, the

\ master shutter on the spectrometer inlet and the condenser charging circuits operate as soon as the starting cam 69 permits. The background shutter remains in position to pass the spectral lines for a full second until moved by the reversing cam, after which the shutter reciprocates at one-second intervals. When the second timer l9 indicates that the condenser charging period should end, the stop cam 81 closes the master shutter and turns off the spark source, both at the end of the next background exposure of the shutter, this proper timing to insure an equal number of line and background" exposures being maintained by the action of the stop cam 8?; independent of the setting of the second timer. As soon as the charging period has ended, the recorder motor starts up. After a halfsecond interval fixed by the measuring cam 90, during which the motor reaches its speed, the measuring cycle is initiated, control of all circuits is transferred to the reference amplifier I relay 2?, and the making of a record on the recorder tape begins. When all condensers are discharged to the predetermined low potential, the recorder motor continues to run out tape for a brief interval set by the delayed-release relay 83, to give enough space to permit the tape to be torn ofi easily, after which all circuits return to their normal state ready for another analysis. The entire operation, which requires thirty to forty seconds to complete an analytical record ready to be interpreted, is thus fully automatic. Errors due to the human element are practically eliminated, and a permanent record, available for checking at any time, is made.

Amplifier A circuit diagram of the direct-current amplifier 25 used in following the falling potentials on the photocell condensers of the spectrometer is given in Fig. 9.

As shown, the first stage of the amplifier consists of a high gain pentode 93 with low screen and plate voltages, e. g. a 6J7 or 606, so that the input resistance will be hi h. This first stage is 81, and a third stage or vpentode 98, e. g. a 6J7, the plate of which is coupled to the second grid of the first stage. The feedback circuit may be broken by the feedback relay switch 51. The terminal of the switch 51 connected through the battery 91 to the control grid of the third stage is also connected to a condenser 99, conveniently .of about 1.0 microfarad capacity, the other side of the condenser being connected to the cathode of the third stage.

As long as the feedback circuit is closed, the amplifier is in a low gain, highly stable state and exhibits little drift. At the same time, the condenser 99 acquires a charge consistent with the balance in the system. Then, when the feedback circuit is opened by the relay 5'! in response to the timing system of the spectrometer, the amplifier is converted to a high gain condition for the measuring period. However, during this period, the condenser 99 tends to retain its charge. preventing the third stage grid potential from changing appreciably during the high-gain period. Thus, when the relay 51 is again closed at the end of the measuring period, the amplifier is returned to the original stable condition. The amplifier of Fig. 9, by virtue of the action of the condenser 99 and the feedback circuit, remains balanced almost indefinitely, and requires adjustment of the voltage-divider 96 perhaps only once aweek while in continuous use.

INTERPRETATION OF RESULTS Fig. 10 illustrates a typical length of the recorder tape of the direct-reading spectrometer direct-coupled to the second stage 94, comprising 1 two triodes in one envelope, e. g, a 6SN'I, operated at conventional voltages and connected to stabilize the plate current. The output of this stage is supplied to the recording pen relay 26, as previously described. The amplifier is rendered highly stable by a feed-back circuit comprising neon glow lamps 95, a voltage-divider 96, battery containing the record of a single analysis of a magnesium-base alloy. The particular record shown was obtained with a spectrometer utilizing eight photocells and gives the analysis of each of seven elements relative to magnesium as the internal reference standard.

In making 'the analysis, the recorder was set to move the tape at a convenient speed, in this case 1.0 inch per second, the duration of the measuring period being about 12 seconds. The recorded line of each element on he tape is made by a pen associated with the photocell, amplifier, and relay corresponding to that element, as already set forth in detail. When the spectrometer reaches its measuring cycle, the recorder first starts the tape moving (from right to left, Fig. 10). Then, as soon as discharge of all condensers begins, the line for magnesium, the reference element, starts to appear. When the potential of the discharging reference condenser reaches that of the standard cell, drawing of all lines ceases and the recorder then stops. The length of the magnesium recorded line thus is a measure of the integrated intensity of the magnesium reference spectral line. The lengths of the other recorder lines are measures of the intensities of the spectral lines of the unknown elements relative to magnesium, and hence measures of the concentrations of those elements, as explained previously.

Fig. 11 is a typical calibration curve forthe unknown element aluminum. The curve was obtained empirically by determining the lengths of the recorder lines observed in sparking electrodes of several different magnesium alloys of known aluminum contents.

For convenience in reading the recorder lines. this calibration curve (Fig. 11) may be projected onto a straight scale (Fig. 12) which is marked 01! to correspond to the aluminum contents as given by the calibration curve. Then, to observe armors the per cent aluminum. as shown by the recorder tape, it is necessary merely to place this prolected calibration scale beside the aluminum re-' corded line, with the zero point of the scale opposite the end of the recorded line, and to read oi! the per cent aluminum from the scale at a point opposite the other end of the recorder line. Similar calibration scales for the other elements shown on the recorder tape may be prepared in like manner, and are used in the same way.

In the event that the calibration curve of an element shifts parallel to itself, due to changes within the optical or electrical-systems of the instrument, it is necessary merely to change the position of the zero point of the projected scale a compensating amount, the magnitude of which may be readily determined byanalyzing a single sample of known analysis. In actual practice, a known sample is run every few hours, and the zero points on the calibration scales of all elements corrected accordingly. In general, the correction thus observed remain exceedingly small,

even over periods of many weeks.

It will be appreciated that the foregoing specification is descriptive, rather than strictly limitatlve, of the present invention, and that numerous variations of the detailed constructions shown are possible without departing from the spirit of the invention, as defined in the claims.

In this specification and claims, the term unknown elementmeans an element the concentration of which is to be determined in the sample being analyzed and the term "reference element" means an element of known concentration in the sample which is being used as an internal standard for the analysis.

What is claimed is:

1. In a direct-reading spectrometer having an exit. slit for a selected spectral line, an electron multiplier photocell for observing the intensity of the line during an exposure period, and'a condenser connected to the output of the cell: a prism mounted at the exit slit in position to divide light passing therethrough into two beams, one beam consisting of the spectral line together with adjacent spectral background and the other consisting of background only, a lens for focusing both beams on the photocell, a shutter between the prism and the photocell provided with means for repeatedly alternating its position at brief equal intervals to cut 01! first one beam and then the other, switching means for repeatedly reversing the polarity of the condenser relative to the photocell in synchronlsm with the alternations of the shutter, the polarities being such that the condenser is charged by the photocell when the shutter is in position to admit the beam containing the spectral line, and means operative after the exposure period for indicating the charge on the condenser.

2. A direct-reading spectrometer for analysis by the internal standard method comprising, in combination with a spark source and a spectrometer: photocells for observing the intensities of the spectral lines of unknown and reference elements; individual condensers for storing the outputs of the photocells during an exposure period;

at brief equal intervals to cut off first one of the beams to each cell and then the other; reversing a master shutter for initiating and terminating the exposure of the photocells to the spectral lines; optical means for focusing on each cell switches for repeatedly reversing the polarity of each condenser relative to its photocell in synchronism with the alternations of the compensating shutter, the polarities being such that the condenser is charged by the photocell when the shutterisinpositiontoadmitthebeam containing the spectral line; and means operative after the exposure for comparing the charges stored in the condensers.

3. In a direct-reading spectrometer comprising a photocell for observing the intensity of a selected spectral line during an exposure period. and a condenser connecwd to the output of the cell, the combination therewith of optical means for focusing on the cell two separate beams, one beam consisting of the spectral line together with its adjacent spectral background and the other consisting of background only, shutter means for alternately interrupting the two beams at brief equal intervals during the exposure period, switching means for ,irepeatedly reversing the polarity of the condenser relative to the photocell in synchronism with the alternations of the shutter. and means operative after the exposure period for indicating the charge on the condenser.

4. In a direct-reading spectrometer comprising a photocell for observing the intensity of a spectral line during an exposure period, and a condenser connected to the output of the cell, the combination therewith of alternating shutter means for repeatedly interrupting and re-establishing at brief intervals of equal length during the exposure period the optical path by which the cell observes the spectral line. switching means for repeatedly reversing the polarity of the condenser relative to the photocell in synchronism with the said alternations of the shutter means. and means operative after the exposure period for indicating the charge on the condenser.

JASON L. BAUNDERSON. VICTOR J. CALDECOURT. EUGENE W. PETERSON.

REFERENCES CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS Number Name Date 1,956,590 Pressler May 1, 1934 2,411,741 Michaelson Nov. 26, 1946 2,436,104 Fisher et al Feb. 17, 1948 FOREIGN PATENTS Number Country Date 144,986 Austria Mar. 25, 1936 OTIER REFERENCES Publication-Long et a1.: Testing Shutter Speeds with the Photo-Electric Cell, page 423 of the Photographic Journal, London, August 1934. (Copy in U. 8. Patent Ofiice Library.)

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US1956590 *Apr 18, 1931May 1, 1934Radio Patents CorpPhoto-electric tube
US2411741 *Dec 13, 1939Nov 26, 1946Gen ElectricColor matching apparatus
US2436104 *May 11, 1944Feb 17, 1948Fisher Scientific CoPhotoelectric apparatus for spectrographic analysis
AT144986B * Title not available
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US2706928 *Jan 6, 1951Apr 26, 1955Lee Foundation For NutritionalSpectrophotometric apparatus
US2855820 *Jun 10, 1955Oct 14, 1958Cons Electrodynamics CorpMethod of spectroscopic gas analysis
US2869364 *Jan 27, 1953Jan 20, 1959Ayres James NApparatus and method for nondestructive testing of initiators
US2972929 *Nov 18, 1952Feb 28, 1961Northrop CorpAutomatic recording photometer
US2999206 *Oct 3, 1958Sep 5, 1961Baird Atomic IncMethod and apparatus for measuring the ratio of electric potentials
US3004465 *Sep 19, 1956Oct 17, 1961White Dev CorpApparatus for reducing effect of scattered light in monochromators
US3024692 *Jun 12, 1958Mar 13, 1962Technicon InstrChart readers and comparators and method of translating light transmittance values of colorimeter recordings into concentration values
US3030854 *May 28, 1957Apr 24, 1962RenaultApparatus for the direct analysis of spectral lines
US3067332 *Sep 9, 1960Dec 4, 1962RenaultDevices for the chemical analysis of materials by means of photocells and by the spectral method
US3102155 *Feb 16, 1955Aug 27, 1963Technicon InstrBackground compensation for spectrometers used in quantitative spectrochemical studies
US3144551 *Mar 28, 1960Aug 11, 1964Atomic Energy Authority UkBeryllium monitor means utilizing an electric arc and spectrum analyzer
US3289777 *Dec 7, 1964Dec 6, 1966Toledo Scale CorpReading system
US3298275 *Apr 19, 1965Jan 17, 1967Cons Electrodynamics CorpComparison signal circuit for measuring intensity of spectral lines
US3439760 *Oct 15, 1963Apr 22, 1969Hobart CorpAutomatic printing price scale with photoelectric encoder including range and motion detectors
US3504978 *Feb 3, 1966Apr 7, 1970Shimadzu CorpPlural beam spectrophotometer with a diffusion plate between each cell and detector
US3533696 *Jul 8, 1966Oct 13, 1970Optische Ind De Oude Delft NvLaser range finder including a light diffusing element
US4074936 *Mar 10, 1976Feb 21, 1978Klockner-Werke AgMethod and apparatus for analyzing a composition of an alloy
US5513002 *Mar 17, 1994Apr 30, 1996The A.R.T. Group, Inc.Optical corona monitoring system
US5550629 *Jun 7, 1995Aug 27, 1996A R T Group IncMethod and apparatus for optically monitoring an electrical generator
US5552880 *Jun 7, 1995Sep 3, 1996A R T Group IncOptical radiation probe
US5764823 *Jun 27, 1997Jun 9, 1998A R T Group IncOptical switch for isolating multiple fiber optic strands
US5886783 *Jun 27, 1997Mar 23, 1999Shapanus; Vincent F.Apparatus for isolating light signals from adjacent fiber optical strands
DE1142711B *Nov 6, 1957Jan 24, 1963Arpad BardoczVerfahren und Einrichtung zur Spektralanalyse mit zeitlich aufgeloesten Spektren ausFunkenentladungen
DE1216580B *Dec 18, 1958May 12, 1966RenaultApparat fuer die chemische Analyse von Werkstoffen im Spektralverfahren
Classifications
U.S. Classification356/306, 250/232, 356/307
International ClassificationG01J3/36, G01J3/30
Cooperative ClassificationG01J3/36
European ClassificationG01J3/36