GB2043878A - Dark Signal Compensation in Spectrophotometers - Google Patents

Abstract

A dual beam spectrophotometer includes signal processing circuitry for producing sample-dark and reference- dark signals. The dark signals subtracted from the sample and ref signals are the average of the dark signals produced on each side of the sample and ref pulses. The signal processing circuitry may comprise four sample and hold circuits 604-607 which store the sample, ref and first and second dark signals, a resistive combining network R630, R631 for the two dark signals and first 608 and second 609 subtractor circuits, the first subtractor 608 gives the sample-dark signal at its output while the second subtractor 609 gives the ref-dark signal at its output. The averaging overcomes problems caused by variations in the background radiation during the measurement cycle, such as can arise in wavelength scanning. <IMAGE>

Description

SPECIFICATION Spectrophotometer The invention relates to dual beam spectrophotometers.

Dual beam spectrophotometers of the ratio recording type, i.e. in which the ratio of the radiation received via a simple path to that received via a reference path is computed to determine the transmittance of a sample substance, include signal processing circuitry for generating from signals produced by a detector a first signal representative of the radiation received by the detector from a radiation source via a first path including a sample cell minus the background radiation and a second signal representative of the radiation received by the detector from the radiation source via a second path including a reference cell minus the background radiation, the radiation received by the detector being in the form of interlaced pulses separated by dark periods in which radiation following the first and second paths is prevented from falling on the detector wherein the detector produces a composite electrical signal representative of the magnitude of the radiation received by the detector.

In such an instrument the radiation received by the detector is in a time division multiplexed form and hence the output of the detector due to background radiation, which is assumed to be the output produced by the detector during the periods during which radiation following the first and second paths is prevented from falling on the detector, is measured at a different time from that during which radiation following the first or second paths is measured. If the magnitude of the background radiation is changing this can give rise to errors in the transmittance value determined. The change in background level can be caused, for example when scanning across the infra red range by selective absorption by watervapour in the atmosphere at different wavelengths.

In prior instruments the background radiation (or dark) level has been measured immediately before the radiation received via the first path (or sample level) is measured and subtracted therefrom to give the first (or sample dark) signal.

The second (or ref-dark) signal is generated in the same way. The detector is normally a.c. coupled to the processing circuitry and hence the sampledark and ref-dark signals may be generated by clamping the signal level to earth, or to some other reference potential, during the periods in which radiation via the first and second paths is prevented from falling on the detector (hereinafter called the dark periods). An alternative known circuit for generating the first and second signals comprises three sample and hold circuits which sample the signal produced by the detector during the periods in which the radiation follows the first path (hereinafter called the sample period), in which the radiation follows the second path (hereinafter called the ref period) and the dark periods respectively.The outputs of the sample and hold circuits are connected to the inputs of two subtractor circuits so that the first substracter circuit produces the sample-dark signal and the second subtractor circuit produces the ref-dark signal. It will be appreciated that with this arrangement, when the level of background radiation is changing the magnitude of the sample-dark and ref-dark signals will be updated during each dark period i.e. during the dark period on each side of the respective sample or ref period. However, neither circuit will prevent dynamic breakthrough onto the first and second signals produced as a result of rapidly changing background radiation levels.

It is an object of the invention to produce a spectrophotometer in which the effect of changes in the level of background radiation on the measurement made is reduced.

The invention provides a dual beam spectrophotometer including signal processing circuitry for generating from signals produced by a detector a first signal representative of the radiation received by the detector from a radiation source via a first path including a sample cell minus the background radiation and a second signal representative of the radiation received by the detector from the radiation source via a second path including a reference cell minus the background radiation, the radiation received by the detector being in the form of interlaced pulses separated by dark periods in which radiation following the first and second paths is prevented from falling on the detector wherein the detector produces a composite electrical signal representative of the magnitude of the radiation received by the detector characterised in that the signal produced by the detector during successive dark periods is averaged and that the average value is subtracted from the signals produced by the detector when the radiation is received via the first and second paths to produce the first and second signals respectively. The average may be taken over two dark periods.

Averaging the signal produced during successive dark periods reduces the effect of changes in background radiation level on the measurernents made and also provides a reduction in the non-linearity of transmittance measurements when the detector is a.c. coupled to the signal processing circuitry or has a low frequency thermal time constant. This applies to typical detectors for infra-red radiation such as pyro-electric or Golay pneumatic detectors.

The signal processing circuitry may comprise four sample and hold circuits, the first storing a signal representative of the radiation received via the first path, the second storing a signal representative of the radiation received via the second path and the third and fourth storing signals representative of the radiation received during successive dark periods, means for averaging the output of the third and fourth sample and hold circuits, a first subtractor having a first input connected to the output of the first sample and hold circuit and a second input connected to the output of the averaging means, and a second subtractor having a first input connected to the output of the second sample and hold circuit and a second input connected to the output of the averaging means, the outputs of the first and second subtractors being the first and second signals respectively.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 shows in diagrammatic form a dual beam spectrophotometer according to the invention, Figure 2 illustrates the response of various processing circuits to a composite signal produces when the level of the background radiation is varying, Figure 3 illustrates the response of certain detectors to the radiation falling on them, Figure 4 is a graph of measured transmittance against actual transmittance when using prior art processing circuitry, Figure 5 is a graph of measured transmittance against actual transmittance when using the circuit of Figure 6, and Figure 6 shows a processing circuit according to the invention for producing output signals representative of the radiation transmitted through the sample and reference cells from the composite signal produced by the detector.

The spectrophotometer shown in Figure 1 comprises a source of radiation S, means for forming two beams of radiation, means for combining the two beams, a monochromator MO, a detector D and signal processing means PC.

Radiation from source S, which may be in the infra-red, visible or ultra violet regions of the spectrum, is reflected by a mirror M1 along the path SB which passes through a sample cell SC within a measurement compartment MC. The radiation following path SB is reflected by two further mirrors M2 and M3 onto a rotating sector mirror assembly M4 which alternately causes the radiation following the path SB to fall on a mirror M8 or to be reflected away from the mirror M8.

Radiation from source S is also reflected by a mirror M5 along a second path RB which passes through a reference cell RC, which is also located in the measurement compartment MC. The radiation following path RB is reflected by two further mirrors M6 and M7 onto the rotating sector mirror assembly M4 which alternately reflects the radiation following the path RB onto the mirror M8 and allows it to pass through and thus be directed away from the mirror M8. Thus a composite beam CB which comprises pulses of radiation which have followed path SB interlaced with pulses of radiation which have followed path RB is formed.The rotating mirror M4 has successive sectors which are radiation transparent, radiation absorbing, radiation reflecting and radiation absorbing in series thus causing the composite beam CB to comprise interlaced pulses of radiation which have followed paths SB and RB respectively separated by periods in which radiation from the source is interrupted. The composite beam CB is reflected by mirror M8 onto an entrance slit SL1 of a monochromator Mo. The monochromator comprises the entrance slit SL1, a concave mirror M9, a diffraction grating G and an exit slit SL2 and is used to select radiation of a narrow band of wavelengths from the wideband radiation presented to the entrance slit SL1.The narrow band radiation emerging from the exit slit SL2 is reflected onto a detector D by a mirror Ml 0. The output of the detector D is fed via signal processing means PC to an indicator I.

In order to determine the transmittance of a sample it is inserted in the measurement compartment MC so as to be traversed by the sample beam and the signal processing means PC is arranged to determine the ratio of the magnitude of the radiation emerging from the sample cell SC to that emerging from the reference cell RC. Since the radiation detected by the detector D during the sample period will be that caused by the radiation from the source which passes through the sample cell plus that due to the background radiation, in order to produce a signal which is due solely to the radiation passing through the sample cell and falling on the detector it is necessary to subtract the dark signal from the sample signal. Similarly the dark signal has to be subtracted from the ref signal before the transmittance, which is the ratio of the sample to the ref signal, can be determined.

Thus the signal processing means PC includes circuitry to perform the necessary subtraction and ratioing.

The signal processing means further includes means for averaging the dark signal produced on both sides of the respective sample or ref signal and uses the averaged signal as the dark signal to be subtracted from the sample or ref signal. This reduces the breakthrough of changing dark signal magnitude onto the sample or ref signals. A further advantage of averaging the dark signals occurring on each side of the sample or ref period is that it reduces the non-linearity of the transmittance signal when the detector is a.c.

coupled to the signal processing means or when the detector itself is an a.c. device.

The waveforms of Figure 2 illustrate the response of the processing circuitry to a composite signal having a negative going drift which may be due, for example, to changing level of background radiation.

With the first type of prior art system discussed hereinbefore the dark signal level is established during the dark periods immediately prior to the respective sample and ref periods and subtracted from the sample and ref signal levels established in the respective periods to produce the sampledark and ref-dark signals. Thus in a period Ta as shown in waveform dthe ref-dark signal will be given by Ra--Dla and the sample-dark signal by Sa-D2a, where Ra is the ref signal level during period Ta, Sa is the sample signal level during period Ta, Dla is the first dark signal level during period Ta and D2a is the second dark signal level during period Ta.However, with the background radiation level drifting as shown in waveform d the true dark signals levels would be (Dla+D2a)/2 and (D2a+Dlb)/2 and thus the true ref signal level would be Ra-(Dla+D2a)/2 and the true sample signal level would be Se- (D2a+D 1 b)/2. With the negative going drift shown in waveform d Dl a will be more positive than (Dla+D2a)/2 and D2a will be more positive than (D2a+Dl b)/2 and hence the ref-dark and sample-dark signals computed will be less than the true values by an amount dependent on the rate of drift of the background radiation level.

Conversely the ref-dark and sample-dark signals will be greater than the true level of the background radiation drifts in a positive direction.

Waveforms e and fillustrate the response of the second type of prior art circuit referred to hereinbefore. In this case the sample-dark and ref-dark signals are not only defined at the end of the ref and sample periods but are modified when each dark signal level is established. Thus for the sample-dark signal illustrated as waveform e during period Tb the sample-dark signal has the value Sa-D2a during the first dark period, Sa Dl b during the ref and second dark periods and Sa-D2b during the sample period. This waveform can be filtered to give a substantially constant output but as can be seen from waveform e it is not symmetrical about the true sample-dark level S-D. The ref-dark signal is shown as waveform fand varies in the same way as the sample-dark signal.Thus with this circuit the computed sample-dark and ref-dark signals will be greater than the true value if the background radiation drifts in a negative direction and smaller than the true value if background radiation drifts in a positive direction.

Waveforms g and h show the response of a processing circuit in which the dark signal level during two successive dark periods is averaged and subtracted from the ref or sample signal. The computed values for the sample-dark signal, shown as waveform g during the period Tb are Sa--(D1 a+D2a)/2 for the first dark period, Sa (D2a+D1 b)/2 for the ref and second dark periods and Sa-(Dlb+D2b)/2 for the sample period. The average value over the period Tb is therefore equal to [Sa-(D1 a+D2a)/2+2[Sa-(D2a+D 1 b)/2]+Sa- (Dl b+D2b)/2]/4=Sa 4(D1b+D2a)/2- 41 (Di a+D2b)/2.

Thus the dark signal level is defined mainly by the average of the dark signal levels during the dark periods on either side of the sample period but with an additional component of one third of the significance due to the dark signal levels next furthest away from the sample period. A similar result applied to the ref-dark signal which is shown as waveform h.

Figure 3 shows a composite waveform which is either a.c. coupled to the processing circuitry or results from a detector such as a pyroelectric or Golay pneumatic detector such as are commoniy used in Infra-Red Spectrophotometers. Figure 3a shows the energy received by the detector and Figure 3b the resulting signal coupled to the processing circuit. The droop orx on each level is proportional to the signal magnitude (x) about its mean level and is given approximately by t"T, where T is the a.c. coupling time constant or the low frequency thermal time constant of an infrared detector, provided that T t.To obtain a good noise filtering characteristic it is necessary to take the mean level of the signal over as long a period as possible, ideally the whole of the ref or sample period, rather than establish instantaneous values. If the measured transmittance is computed as (Sa-D2a)/(Ra-D1 a), where Sa, Ra, Dla, D2a are the mean signal levels during the respective sample reference and dark periods, as will be the case with the first prior art circuit and if crA a few % then the measured transmittance will be given by Tm Ta+a/4(1-- Ta2), where Tm is the measured transmittance and Ta is the actual transmittance.

As can be seen from Figure 4 this will give a zero offset of a/4 and a maximum deviation from linearity of a/i 6 when Tea=3. If a=5% then a/4=1.25% and all 6=0.31%. If If however the measured transmittance is computed as Sa-(D1 b+D2a)/2 Ra--(D a+D2a)/2 then the measured transmittance will be given by Tm=Ta +(1 -Th2)a2/S.

As can be seen from Figure 5 this gives a zero offset of a2/8 and a maximum deviation from linearity of cm2/32. If a=5% this gives a zero offset of 0.31% and a maximum deviation from linearity of 0.0078%.

Thus a significant improvement in linearity and in the offset of the measured transmittance from the actual value can be achieved by averaging the dark signal levels during the dark periods on each side of the respective sample and reference periods.

Figure 6 illustrated processing circuitry in which this averaging may be performed. The processing circuitry comprises four sample and hold circuits, 604, 605, 606 and 607 and two subtractor circuits 608 and 609.

Sample and hold circuit 604 comprises an FET T604, a resistor R604, a capacitor C604 and an operational amplifier A604, sample and hold circuit 605 comprises an FETT605, a resistor R605, a capacitor C605 and an operational amplifier A605, sample and hold circuit 606 comprises an FET T606, a resistor R606, a capacitor C606 and an operational amplifier A606 and sample and hold circuit 607 comprises an FET T607, a resistor R607, a capacitor C607 and an operational amplifier A607.

The subtractor circuits 608 and 609 are identical in form, subtractor 608 comprising an operational amplifier A608 having its positive input biased by a resistor network comprising resistors R608, R609, R610 and R61 1, a capacitor C608 being connected across resistors R608 and R609. The parallel arrangement of a resistor R6 12 and a capacitor C6 1 2 is connected between the output and the negative input of amplifier A608. A first input of subtractor 608 is connected via a resistor R61 3 to the negative input of amplifier A608 while a second input is connected via a resistor R6 14 to the positive input of amplifier A608.Subtractor 609 is identical in form and comprises resistors R6 1 8- R623, capacitors C618 and C622 and operational amplifier A609.

The composite waveform is applied on line 601 to the sample and hold circuits 604-607.

Timing signals are fed from terminals 613, 614, 615 and 616 respectively to sample and hold circuits 604-607. In this way during the first dark period sample and hold circuit 607 samples the composite waveform, as the timing signal causes FET T607 to conduct, and stores the magnitude of the composite waveform at that time on capacitor C607. In this way a representation of the radiation falling on the detector D during the first Dark period is stored. In the same way value of the Reference signal is stored on capacitor C606, the sample signal on C604 and the signal during the second Dark period on capacitor C605.

The output of sample and hold circuit 604 is connected to a first input of the subtractor circuit 608 while the output of sample and hold circuit 606 is connected to a first input of a subtractor circuit 609. The output of sample and hold circuit 605 is connected to one end of a resistor R630 while the output of sample and hold circuit 607 is connected to one end of a resistor R63 1. The other ends of resistors R630 and R63 1 are commoned and connected to second inputs of subtractors 608 and 609.

Thus subtractor 608 has the latest sample signal applied to its first input and the average of the last two dark signals applied to its second input while subtractor 609 has the latest ref signal applied to its first input and the average of the last two dark signals applied to its second input. The output of subtractor 608 on terminal 617 will consequently be as shown in Figure 2 as waveform g and that of subtractor 609 on terminal 618 that shown as waveform h.

Appropriate filtering can be applied to these waveforms and the ratio taken to produce the transmittance value. Circuitry suitable for obtaining this ratio and details of the timing signals are disclosed in more detail in our copending application No. 7907618. Serial No.

2043880.

It would be possible to average the dark signal level over a greater number of dark periods by providing additional sample and hold circuits and further combining networks. It would, of course, be necessary to generate further timing signals to gate the composite signal to the appropriate sample and hold circuit.

In an alternative embodiment the composite signal may be sampled by an integrating digital voltmeter and stored as binary digits. The sampledark and ref-dark signals can then be computed in the arithmetic unit of a central processor and the signals representing the dark levels averaged over any desired number of cycles. It would then be possible to weight the significance given to the measured dark signal magnitude to take into account the effect of changes other than in the immediately adjacent dark periods.

Claims (5)

Claims

1. A dual beam spectrophotometer including signal processing circuitry for generating from signals produced by a detector a first signal representative of the radiation received by the detector from a radiation source via a first path including a sample cell minus the background radiation and a second signal representative of the radiation received by the detector from the radiation source via a second path including a reference cell minus the background radiation, the radiation received by the detector being in the form of interlaced pulses separated by dark periods in which radiation following the first and second paths is prevented from falling on the detector wherein the detector produces a composite electrical signal representative of the magnitude of the radiation received by the detector characterised in that the signal produced by the detector during successive dark periods is averaged and that the average value is subtracted from the signals produced by the detector when the radiation is received via the first and second paths to produce the first and second signals respectively.

2. A dual beam spectrophotometer as claimed in Claim 2 in which the average is taken over two dark periods.

3. A spectrophotometer as claimed in Claim 2 in which the signal processing circuitry comprises four sample and hold circuits, the first storing a signal representative of the radiation received via the first path, the second storing a signal representative of the radiation received via the second path and the third and fourth storing signals representative of the radiation received during successive dark periods, means for averaging the output of the hird and fourth sample and hold circuits, a first subtractor having a first input connected to the output of the first sample and hold circuit and a second input connected to the output of the averaging means, and a second subtractor having a first input connected to the output of the second sample and hold circuit and a second input connected to the output of the averaging means, the outputs of the first and second subtractors being the first and second signals respectively.

4. A spectrophotometer as claimed in Claim 3 in which the averaging means comprised a resistor network.

5. A dual beam spectrophotometer substantially as described herein with reference to Figures 1 to 5 or to Figures 1 to 6 of the accompanying drawings.