Spectrophotometers 15 value at a specific the memory and by retrieving it from the memory for comparison with the extinction values wavelength Several manufacturers have taken advantage of this possibility in their software. Diode-array spectrometers are an interesting new development pioneered by Hewlet-Packard.Instead of a monochromator,a polychromator is used in which the dispersed light of a continuum source is brought to a focus in one plane,the focal plane of the instrument.This method of operation cor- responds to the classical method where a photographic ate which 'recorded"the whole spectrum directly,was fitted in the focal plane of a -diode ar de ctor is used.instead of a fast electronic pro ocessing of the spectral information stored for a sh e in of 256 or 512 diodes (channels).The comp re not arrange as In trophotometer,in particular the cuvettes are mounte ed conver the ligh source and the entry slit of the polychro Figure6 shows the optical layout of the Perkin-Elmer Lambda-Array UV-VIS Spectrometer 3840.The light of a deuterium lamp is focusec 1 on to the sample by the plane(flat)and toroidal mirrors(M2 and M3)and,after traversing the sample,arrives via mirror Ma and another plane mirror at the entry slit of the polychromator.Here the continuum of the light source is directed onto a grating via Ms.Finally the light,now spectrally dispers- ed,falls via ror Ms onto the photodiode array(PDA)detector with 512 grating mounted on a turntable.The first grating use for the fast quisition of a survey spectrum in the region (th olution 1 5n effective resolution ca. 4.5 nm,i.e the survey mode). Th (600 line Vmm)is used in a high-performance mod ch eps achiev resolution(≈0.25or ≈0.75nm).In spectr ons 100 nm arrive at the diode array,i.e.the wh region of divided into 7 sections.Depending on the width of the selected region the time for measuring a whole spectrum is approximately 8 16s.Th deuterium lamp (type D 802 or D902,Heraeus Original Hanau)transmits light along the optical axis and,as shown in Fig.6,this allows the introduc tion of the beam of a tungsten lamp into the light path.The W and source filter wheels are used to cut out one lamp when measuring with the other All functions(filter change,shutter,rotation of grating,PDA read-out,col- lecti of etc.)are ontrolled by the Perkin-Elmer PC7500. ed a s imilar instrument,model HP 8452A. Philips and Milton Roy also market array spectrometers
Spectrophotometers 15 value at a specific wavelength Ai into the memory and. by retrieving it from the memory for comparison with the extinction value's at wavelengths A2' A3 . . . . Several manufacturers have taken advantage of this possibility in . their software. Diode-array spectrometers are an interesting new development pioneered by Hewlet-Packard. Instead of a monochromator, a polychromator is used in which the dispersed light of a continuum source is brought to a focus in one plane, the focal plane of the instrument. This method of operation corresponds to the classical method where a photographic plate, which "recorded" the whole spectrum directly, was fitted in the focal plane of a spectrograph. Today a silicon-diode array detector is used, instead of a photographic plate, and this allows direct, fast electronic processing of the spectral information stored for a short time in one row of 256 or 512 diodes (channels). The components are not arranged as in a conventional spectrophotometer, in particular the cuvettes are mounted between the light source and the entry slit of the polychromator. Figure 6 shows the optical layout of the Perkin-Elmer Lambda-ArrayUV-VIS Spectrometer 3840. The light of a deuterium lamp is focused on to the sample by the plane (flat) and toroidal mirrors (M2 and M3) and, after traversing the sample, arrives via mirror M4 and another plane mirror at the entry slit of the polychromator. Here the continuum of the light source is directed onto a grating via Ms. Finally the light, now spectrally dispersed, falls via mirror M6 onto the photo diode array (PDA) detector with 512 photodiodes. Two gratings are mounted on a turntable. The first grating (100 lines/mm) is used for the fast acquisition of a survey spectrum in the 200-900 nm region (theoretical resolution 1.5 nm, effective resolution ca. 4.5 nm, i.e. the survey mode). The second grating (600 lines/mm) is used in a high-performance mode in which it is rotated in 7 steps to achieve a high resolution (., 0.25 or ., 0.75 nm). In this mode, spectral regions of ca. 100 nm arrive at the diode array, i.e. the whole region of measurement is divided into 7 sections. Depending on the width of the selected A region, the time for measuring a whole spectrum is approximately 8 - 16 s. The deuterium lamp (type D 802 or D 902, Heraeus Original Hanau) transmits light along the optical axis and, as shown in Fig. 6, this allows the introduction of the beam of a tungsten lamp into the light path. The Wand source filter wheels are used to cut out one lamp when measuring with the other. All functions (filter change, shutter, rotation of grating, PDA read-out, collection of data, etc.) are controlled by the Perkin-Elmer PC 7500. Hewlett-Packard have developed a similar instrument, model HP 8452A. Philips and Milton Roy also market diode array spectrometers
MI ',kRWHm W LAMP / . -.- % GtLAM' rl-i?FLAT M 2 u-=::-_ APERTURES I _.-"U FLAT M3\r -+- SOURCE FILTER WHEEL Fig. 6. Lambda array 3840 optical layout 0"1 - I PDA 0- M6 /- TURNTABLE 2- i I 'U M5 SAMPLE I _ AM4 +1 , I FLAT \r
The Stray Light Error 17 3.3 The Stray Light Error 3.3.1 General Observations Since the vast majority of UVVIS still fitted with a single monochromator,rrors to le4g0enatenoceoaihoaegA320 >45000 cm);particularly in the case of small transmission values. By stray light we mean light of other wavelengths which is superimposed upon the useful light.Often mistakenly called scattered light,it is caused by scattering at the optical surfaces in the monochromator. If a monochromator is set to wavelength 4o and if the slitwidth cor- dsto the effective bandwidh the usefu light lies in the eion 0-41ando+41 (8) sion decreasing linearly from Ao in both directions.However,due to stray light there is a certain transmission by the monochromator of light outside the useful light region.Although its intensity is small (order of magnitude 10-5)it can still be significant since the radiation detector sums the stray light over the whole wavelength region to which it is sensitive and in which the light source emits radiation. The proportion of the photoelectric current given out by the detector and the display.which is due to stray light,is the critical measure of when making photoe e t the term ortion of stray light the nt arising from tray photoelectric the photo arising from the stray light is generally low (<0.1),it can ca cant problems if the photoelectric current due to the useful light becomes relatively small.In practice,this occurs in the following cases: 1.The useful light can be weakened by absorption occurring in the light theh is hardly diminished.Thisr rtic a)The optical elements in the light path (envelope of the hydrogen lamp. source mirror,lenses,mirrors,prisms or gratings in the monochromator lenses in the sample changer,multiplier envelope)absorb increasingly with decreasing wavelength.In addition,there is the effect of the pres- ence of absorbing deposits on the accessible optical surfaces (envelope of the hydrogen lamp,lenses in the sample changer,cuvettes).Further- more,atmospheric oxygen which is present in the optical path of the in- strument absorbs at wavelengths below 200nm
The Stray Light Error 17 3.3.1 General Observations Since the vast majority of UV-VIS spectrophotometers are still fitted with a single monochromator, errors can occur due to monochromator stray light at the limits of the monochromator transmission (e.g. A :;;:;220 nm .!;. ~ 45 000 cm -1); particularly in the case of small transmission values. By stray light we mean light of other wavelengths which is superimposed upon the useful light. Often mistakenly called scattered light, it is caused by scattering at the optical surfaces in the monochromator. If a monochrom::ttor is set to wavelength AO and if the slitwidth corresponds to the effective bandwidth LlA, the useful light lies in the region between (8) We shall call this the useful-light region. In ideal circumstances, a monochromator should only transmit in the useful light region with the transmission decreasing linearly from AO in both directions. However, due to stray light there is a certain transmission by the monochromator of light outside the useful light region. Although its intensity is small (order of magnitude 10-5) it can still be significant since the radiation detector sums the stray light over the whole wavelength region to which it is sensitive and in which the light source emits radiation. The proportion of the photoelectric current given out by the detector and reaching the display, which is due to stray light, is the critical measure of the stray light effect when making spectrophotometric measurements. For that reason, we understand by the term proportion of stray light the ratio of the photoelectric current arising from stray light to the total photoelectric current. Although the proportion of the photoelectric current arising from the stray light is generally low ( < 0.1 070), it can cause significant problems if the photoelectric current due to the useful light becomes relatively small. In practice, this occurs in the following cases: 1. The useful light can be weakened by absorption occurring in the light path whilst the stray light is hardly diminished. This occurs particularly below 230 nm because: a) The optical elements in the light path (envelope of the hydrogen lamp, source mirror, lenses, mirrors, prisms or gratings in the monochromator, lenses in the sample changer, multiplier envelope) absorb increasingly with decreasing wavelength. In addition, there is the effect of the presence of absorbing deposits on the accessible optical surfaces (envelope of the hydrogen lamp, lenses in the sample changer, cuvettes). Furthermore, atmospheric oxygen which is present in the optical path of the instrument absorbs at wavelengths below 200 nm
18 Photometers and Spectrophotometers b)If the solvent absorbs in the short-wavelength UV region but the long wavelength stray light is transmitted undiminished,the proportion of stray light therefore increases.Special attention should be paid to this possibility since most solvents absorb in the short-wavelength UV region due to the presence of impurities if they have not been especially purified.For many solvents,preferred because of their solvent proper- ties,the transmission at 40000 cm-1(=250 nm)is practically zero, therefore.for these solvents.stray light must be taken into consideration below 260 to 270 nm. 2.In some sp t region is relati he inte measuring fore,stray-light protective filters,which absorb the long-wavelength stray light,should be used in this region.These filters can reduce the propor- tion of stray light to less than 0.2%for measurements made with solvents free from the problems mentioned under 1 b. 3.In certain spectral ranges,the detector sensitivity in the useful-light mall cor d with that in the stray light regic case t the lo ng- length limit of the der ulti e appro ma ey620 nd for r photoc bove 1μm r that mea exte beyond these. the a consi error unless th short-wavelength stray light is reduced by special filters. Case 1 is particularly important in practice and requires careful control of the proportion of stray light during measurement if the solvent shows considerable absorption in the useful-light region.In any event,the absor- bance of the solvent should be tested when measuring below 230 nm.For this purpose,the absorbance of the reference cuvette can be measured against air.If the absorbance of the solvent is greater than approximately 0.5,t or po the absorbance either by purifying the sol. reduc ld be in nves clear a stray light erro 人 ng solver s manifests is made with ath ths and the wa ngth e sar value of the extinction coefficient should be obtained for all pathlengths This is often the case,within the error threshold,for wavelengths above ca 250 nm.In contrast,in the presence of a stray light error,the curves diverge with decreasing wavelength in the sense that smaller values of the extinction coefficient are obtained for longer pathlengths.Obviously,values obtained at the shortest pathlength are the most reliable ones
18 Photometers and Spectrophotometers b) If the solvent absorbs in the short-wavelength UV region but the longwavelength stray light is transmitted undiminished, the proportion of stray light therefore increases. Special attention should be paid to this possibility since most solvents absorb in the short-wavelength UV region due to the presence of impurities if they have not been especially purified. For many solvents, preferred because of their solvent properties, the transmission at 40000 cm -1 (= 250 nm) is practically zero, therefore, for these solvents, stray light must be taken into consideration below 260 to 270 nm. 2. In some spectral ranges, the intensity of the radiation from a light source in the useful-light region is relatively small in relation to the intensity in the region generating the stray light. This is the case when measuring with an incandescent lamp in the range between 320 and 400 nm. Therefore, stray-light protective filters, which absorb the long-wavelength stray light, should be used in thibS region. These filters can reduce the proportion of stray light to less than 0.20/0 for measurements made with solvents free from the. problems mentioned under 1 b. 3. In certain spectral ranges, the detector sensitivity in the useful-light region is relatively sm,!illl compared with that in the stray light region. This is the case at the long-wavelength limit of the detector sensitivity, i.e. for photomultipliers above approximately 620 nm and for photocells above 1.1 J.1m. For that reason, measurements cannot be extended beyond these limits without the risk of a considerable error unless the short-wavelength stray light is reduced by special filters. Case 1 is particularly important in practice and requires careful control of the proportion of stray light during measurement if the solvent shows considerable absorption in the useful-light region. In any event, the absorbance of the solvent should be tested when measuring below 230 nm. For this purpose, the absorbance of the reference cuvette can be measured against air. If the absorbance of the solvent is greater than approximately 0.5, the possibility of reducing the absorbance either by purifying the solvent or by reducing the pathlength should be investigated. A clear indication of a straY light error by absorbing solvents manifests itself as follows. If the measurement is made with different pathlengths and the extinction coefficient recorded as a function of the wavelength, the same value of the extinction coefficient should be obtained for all pathlengths. This is often the case, within the error threshold, for wavelengths above ca. 250 nm. In contrast, in the presence of a stray light error, the curves diverge with decreasing wavelength in the sense that smaller values of the extinction coefficient are obtained for longer pathlengths. Obviously, values obtained at the shortest pathlength are the most reliable ones
The Stray Light Error 19 3.3.2 The Stray Light Error of Transmission and Absorbance and Its Measurement The useful light leaving the reference cuvette induces the ph otoelectric cur rent Io in the detector.The photoelectric current generated by the useful light leaving the sample cuvette is I and the real transmission (T)of the sam ple is T=I (9) The stray light induces the additional photoelectric current Is.Conse- quently,the measurement gives a false transmission T'which has the value T'= I+Ig Io+Ig (10) This assumes,in the first instance,that the stray light is reduced by the sample in the same manner as by the solvent.This assumption is adequate in many cases. On introducing the proportion of stray light,P. I。 p= Io+Ig we obtain T'=T(1-p)+p (11) If the proportion of stray light is known,the true transmission can be calculated from the value distorted by the stray light as T=T-p 1-p (12) When moving from transmission to absorbance the following equations A=-logT A=-logT provide the absorbance error caused by stray light 4A =A'-A log T-log [T(1-p)+p]. (13) The relative error A'of the absorbance is of particular interest in spec- trophotometry.This value is plotted in Fig.7 as a function of the distort absorbance Awhich is read directly [7].The proportion of stray light,P. has been included as a show error in the absorbance in crease rapid. given proportion o stray light p ngth an concentration must be selected for all mea surements such that the absorbance is not too large.At p values of the order
The Stray Light Error 19 3.3.2 The Stray Light Error of Transmission and Absorbance and Its Measurement The useful light leaving the reference cuvette induces the photoelectric current 10 in the detector. The photoelectric current generated by the useful light leaving the sample cuvette is I and the real transmission (T) of the sample is I T=-. 10 (9) The stray light induces the additional photoelectric current If. Consequently, the measurement gives a false transmission T' which has the value T' = 1+lf . 10+lf (10) This assumes, in the first instance, that the stray light is reduced by the sample in the same manner as by the solvent. This assumption is adequate in many cases. On introducing the proportion of stray light, p, If p=- 10+lf we obtain T' = T(l-p)+p (11) If the proportion of stray light is known, the true transmission can be calculated from the value distorted by the stray light as T'-p T=-. 1-p (12) When moving from transmission to absorbance the following equations A = -log T; A' = -log T' provide the absorbance error caused by stray light LlA = A'-A = log T-Iog [T(l-p)+p] . (13) The relative error LlA' of the absorbance is of particular interest in spectrophotometry. This value is plotted in Fig. 7 as a function of the distorted absorbance A' which is read directly [7]. The proportion of stray light, p, has been included as a parameter. The graph shows that the relative error in the absorbance increases rapidly with the value of the absorbance for a given proportion of stray light p. Therefore, the pathlength and concentration must be selected for all measurements such that the absorbance is not too large. At p values of the order