Anal. Chem. 1988, 6 0 , 2258-2263
2258
DL (eq 2) was found to be 5 X mmol (2 ng) in the case where a suppressor was used and between 2 X lo-' and 5 X lo-' mmol(8 and 20 ng) without a suppressor, depending on the eluent used. nmincalculated from DS (eq 1) was found to be in the range 4 X to 7 X mmol (150-250 pg). The value of DL, calculated by eq 8 or eq 14 from DS characteristic of the detector and from column parameters is in good agreement with DL calculated by eq 2 from actual chromatograms, which indicates that the value of DL, most important analytically, can be predicted from knowledge of the detector sensitivity and column parameters. Registry No. I-, 20461-54-5;NO3-, 14797-55-8;Br-, 24959-67-9; C1-, 16887-00-6;CH3COO-, 71-50-1; Br09-, 15541-45-4.
LITERATURE CITED (1) Tesarlk, K.; Kallb, P. J. Chromatogr. 1973, 78, 357-361. (2) Svobcda, V.; Marsal, J. J . Chromatogr. 1978, 148, 111-116.
Molnlr, I.; Knauer, H.; Wiik, D. J. Chromatogr. 1980, 201, 225-240. Johnson, D. E.; Enke, C. G. Anal. Chem. 1970, 42, 329-334. Jackson, P. E.; Haddad, P. R. J . Chromatogr. 1986, 355, 87-97. Glatz. J. A.: Girard, J. E. J. Chromtogr. Sci. 1982, 20,266-273. Jenke, D. R.; Pagenkopf, G. K. Anal. Chem. 1982, 54, 2803-2604. Cassidy, R. M.; Elchuk, S. J. Chromatcgr. Sc;. I983#21, 454-459. Pungor, E.; Pli, F.; T6th, K. Anal. Chem. 1883, 55, 1728-1731. Amati, D.; sz. KovBts, E. Langmuir 1987, 3, 887. Gobet, J.; sz. Kovlts, E. Adsorpt. Sci. Technol. 1984, 1 , 77. T6th, P.; Kugler, E.; sz. Kovlts, E. Helv. Chim. Acta 1959, 42, 2519-2530. Gustafson, F. J.; Markell, C. G.; Simpson, S. M. Anal. Chem. 1985, 57,621-624. Jenke, B. Anal. Chem. 1981, 53, 1535-1536. Okada, T.; Kuvamoto, T. Anal. Chem. 1983, 55, 1001-1004. Gjerde, D.T.; Fritz, J. S. Anal. Chem. 1981, 53,2324-2327. Okada, T.; Kuvamoto, T. Anal. Chem. 1984, 56, 2073-2078.
RECEIVED for review October 14, 1987. Accepted June 13, 1988. P. F. gratefully acknowledges the financial help of the Hamilton Foundation (S.E.A.).
Acoustic Signal as an Internal Standard for Quantitation in Laser-Generated Plumes Guoying Chen and Edward 5.Yeung* Ames Laboratory-USDOE and Department of Chemistry, Iowa State University, Ames, Iowa 50011
To correct for puke-te.puke v(rrktlww In lasergenerated plumes, one needs to monltoT th. total mount of materlal vaporized durlng each puke. We f p M that the magnitude of the acoustic wave arsoalated with @utwgeneration Is linearly related to the OmWon lntenritles of both major and minor elements In the rl(l Over L wkle range of vaporkatbn pretrea signals the acoustlc slgnal can be used as an lnterw3I standard for normallzlng analytlcal signals derived frdm lasergenerated plumes.
The laser microprobe analyzer (LMA) was originally developed to enable qualitative analysis with high spatial resolution for samples of diverse natures (1). LMA has established itself as an appropriate tool for this purpose. Efforts have been made to improve its reproducibility so that it can be used as a reliable quantitative tool. However, so far this goal has not been fully realized. The reason is multifaceted. Firstly, as expected, the analytical precision is highly dependent on fluctuations in the laser power on a shot-to-shot basis (2). Secondly, matrix effects contribute to the irreproducibility of the result of analysis. This includes two factors, namely physical and chemical matrix effeds (3). The former refers to the influence exerted by the mechanical, physical, and also chemical nature of the sample surface on the vaporization process. Examples are grain size ( 4 ) , mechanical tension (4),crystal orientation ( 5 ) ,chemical composition (6),and so on. The latter, on the other hand, refers to the effects of foreign elements on the chemical composition, the total electron density, and the subsequent excitation process in the plasma, and consequently 0003-2700/88/0360-2258$01.50/0
on the spectroscopicsignals. This effect is closely tied to the composition of the solid sample itself (7). Finally, variation in experimental conditions is another factor to consider. These include laser focusing, angle of incidence of the vaporization laser, surface condition changes due to etching, different sample treatment procedures, oxidation conditions, contamination, and so on. Unfortunately, these factors generally cannot be precisely controlled. So, LMA remains only a qualitative or semiquantitative tool (3, 8), with a relative standard deviation usually in the range of 10-30%. Improved precision can be obtained by signal averaging over a number of laser pulses (9). However, the cost here is throughput and spatial resolution, since as different spots or different depths are being sampled, the physical conditions and chemical compositions might not be identical. This approach is thus not applicable to inhomogeneous solid samples. Other alternatives have been suggested to improve the analytical precision. These include the use of standard material with closely matched composition (IO),monitoring the power fluctuations of the vaporization laser ( I I ) , monitoring the size of the crater on the sample surface produced by the laser shot (12),and the use of internal standards (13). Among these methods, the use of standard material is highly recommended to improve the analytical accuracy (14). But, this cannot take into account the variations of experimental parameters, including laser power fluctuations, so the improvement in precision is quite limited. Besides, it is not always possible to find a standard sample close enough in composition to that of the sample under study. Monitoring the laser power results in, unfortunately, no improvement in precision of analysis (11, 15),since higher power does not necessarily produce a higher signal intensity. Instead, it can generate more ions, especially multiply charged and more energetic ions (16). Besides, reflectivity and thus absorption of the target material is intensity dependent. Crater size measurement is extremely tedious and 0 1988 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 60, NO. 20, OCTOBER 15, 1988
cannot be done in situ (17). The minute amounts of material removed by a single laser shot are often difficult to measure, and large errors result. Moreover, this cannot take into account those droplets and clusters redepositing onto the sample or the cell wall, which are removed from the crater but do not contribute to the analytical signal. The most useful technique among these is probably the use of an internal standard, i.e., measurement of analyte-tostandard intensity ratio instead of the analyte line intensity itself. The assumption here is that interferences affecting the intensities of these two lines are identical. Although this is not always the case, as observed by Margoshes and co-workers (I&?), intensity ratio measurement in many cases does improve the reproducibility considerably (13,14). Another assumption behind this method is that the standard element selected is uniformly distributed over the volume of interest. This condition, however, cannot always be met. For instance, rocks and inclusions represent highly inhomogeneous cases. Depth profiling is another example, where the surface layer might have quite a different composition compared to that of the bulk. On these occasions, other methods of normalization are definitely needed. Since one of the useful characteristics of laser sampling is its relative freedom from evaporation selectivity (19, ZO),the basic question is to choose a measurement that correlates more directly to the amount of material removed by each laser shot so that it can be used to normalize the intensity of the spectroscopic signal to yield better reproducibility. The density probe based on beam deflection, capable of monitoring the amount of material removed by individual laser pulses (21), has the potential to serve this purpose. We report in the following another normalization technique based on monitoring the acoustic signal associated with the plume generated by pulsed laser radiation. The acoustic wave is detected by a microphone located inside the sample cell, into which a buffer gas a t a selected pressure was introduced. Under widely varying experimental conditions, the amplitude of the acoustic wave is compared to the emission intensities of both minor and major components in the sample. The results indicate that the precision of laser microprobe sampling can be improved by using the acoustic wave as an internal standard.
EXPERIMENTAL SECTION A block diagram of the instrumental arrangement is shown in Figure 1. A pulsed excimer laser (Lumonics, Ottawa, Canada, Model Hyper EX 460) running at the 308-nm XeCl transition was used as the vaporization laser. The UV beam was directed to the sample cell with a mirror (Newport, Fountain Valley, CA, 10QM20EM.15) and focused onto the sample surface by using a planoconvex UV grade quartz lens with 2-in. diameter and 1Wmm focal length (Oriel,Stratford, CT). This lens was mounted on a micrometer translational positioner so that the focusing condition could be finely adjusted. The pulse energy used was regulated by varying the operating voltage and using variable apertures. The energy of each individual pulse was monitored by an energy ratiometer (Laser Precision, Utica, NY, Model Rj-7200) with an energy probe (Laser Precision, Utica, NY, Model RjP 734). The pulse duration was 25 ns, and the repetition rate was typically 1Hz. Typical energy used in this work was less than 6 mJ/pulse. The pulse-to-pulseenergy reproducibility was found to be h1.4%. The solid samples were cut into l/&. X l/s-in. X 1/16-in.pieces and then mounted on the 45O inclined end surface of a stainless steel rod, which was screwed directly to a homemade micrometer 3D/tilt stage. In this work severalpure metallic and alloy samples were used, including NBS SRM 1222 Cr-Ni-Mo steel, SRM C1150a white cast iron, and SRM C1288 high-alloy steel (National Bureau of Standards, Washington,DC). The sample surfaceswere polished with No. 600 gritpaper. Some were further polished on a Buehler polisher/grinder (Buehler, Lake Bluff, IL, Model polishingpowders grade A and/or B with Ecomat 111) using A1203
SIDE V I E W
i TOP V I E W
2259 I
t Flgure 1. Experimental arrangement for simultaneous measurement of emission and acoustic wave assoclated with lasergenerated plume: A, aperture; C, sample cell; HV, highvoltage power supply; L1 and L2, UV grade pianoconvex quartz lenses; M, mirror; MC, monochromator; MP, condenser microphone; PA, preamplifier; SA, sample piece; SL, vaporization laser; ST, 3D/tilt micrometer stage; WA, waveform ana-
lyzer. particle diameters of 0.3 and 0.05 micrometers, respectively. The sample surfaces were cleaned with organic solvents before the experiments. To ensure position reproducibility of the sample, a position reference bar defined by two precision pinhole pairs was used. A new sample piece was glued on by using Bmin epoxy against a 45' groove in this bar and the original position recorded. A Pyrex cell was mounted from the front against an O-ring to ensure airtightness. The cell was a cross in shape, made from 2 cm diameter Pyrex tubing. Quartz windows were attached to the ends of the tubing to introduce the pulsed UV laser beam and to monitor the emission. On the bottom of the cell a 0.5 in. diameter condenser microphone (Knowles, Franklin Park, IL, Model BT-1759) was mounted to measure the acoustic wave generated by the laser plume. This type of microphone was used because of its relatively flat frequency response, and hence suitability for measuring impulse signals, and ita insensitivity to mechanical vibrations. The cell volume (about 30 cm3)was not minimized so that deposition of the ablated material on the microphone diaphragm and on the windows could be avoided. The top of the cell was connected to a vacuum chamber so that a buffer gas at selected pressures could be introduced. A two-stage glass diffusion pump was used in the vacuum system. The chamber pressure was measured with a capacitance vacuum gauge (MKS Instruments, Burlington, MA, Type 221AHS-F-1000). Helium at 50 Torr was used as the buffer gas. In this work, a spatial masking scheme was employed. As shown in Figure 1,the continuum background emission from the crater area was blocked by the sample piece itself, which was tilted at 45O. Only the emission from the top periphery of the plume was collected by a planoconvex UV grade quartz lens with 1-in. diameter and 3.5-cm focal length (Oriel,Stratford, CT) and imaged onto the entrance slit of a monochromator (PTR Optics, Waltham, MA, Model ptr MC1-03). The monochromatorhad an f number of 4 and a reciprocal linear dispersion of 6 nm/mm. The wavelength reading was calibrated by using a HeNe laser and several pure metallic samples. The slit width used in most cases was 150 micrometers. The emission intensity was recorded by a photomultiplier tube (PMT) (Hamamatsu, Bridgewater, NJ, Model R928). Because of the reproducibilityof the sample in space, both masking conditions and light collection efficiency could be kept fairly constant from sample to sample and from run to run. All the optical components were mounted on a 4-ft. X 6-ft. X 2-in. optical table (Newport, Fountain Valley, CA, Model NRCXS-46) padded with a in. thick rubber sheet. Both signals from the PMT and microphone were sent to a two-channel waveform analyzer (Data Precision, Danvers, MA, Model D6W) with a two-channel preamplifier (Data Precision, Danvers, MA, Model D1000). The waveform analyzer had an 8-bit amplitude
2260
ANALYTICAL CHEMISTRY. VOL. 60, NO. 20, OCTOBER 15. 1988
Temporal dependence of continuous (plasma)emission (len peak) and line (atomic)emission (shouiders at the right). The star? of the trace is synchronized with the laser pulse. Conditions: sample. NES SRM 1222 Cr-Ni-Mo steel; laser, 308 nm. 25 ns, 0.95 W: He at 50 Torr: 40 nslpoint; No. 4-96 filter without spatial mask. Flgurs 2.
resolution and a sampling rate of 10 ns/point for one channel or 20 ns/point for two channels. Alternating current coupling mode was selected for all signal channels to reject the long-term noise. The peak area and height as well as other parameten are calculated by using the mathematical functions build into the waveform analyzer. Linear regression was done on an IBM PC/AT personal computer.
RESULTS AND DISCUSSION Optimization of Conditions. Laser sampling is often accompanied by plasma emission (22). While it might not be a problem in maw spectrometry and in cases where the plume is swept by an inert gas stream to another location for analpis, it causes severe interference to in situ optical spectroscopic determinations. To minimize this continuous background, several approaches were taken in this work. (1) A UV laser a t 308 nm was used as a vaporization laser. In laser mass spectrometry, the wavelength of the laser is not a primary parameter (23). I t is relevant only because most metals absorb strongly in the UV range, resulting in effective evaporation (24)and a shallow skin depth, which is desirable in depth profiling. In optical spectmscopy, since the coefficient of inverse bremsstrahlung absorption, the major mechanism for plasma absorption, decreases as the 2nd to 3rd power of the wavelength of the absorbed photons (%), short wavelength radiation minimizesplasma generation and muples better with solid samples than long wavelengths. A reduced background continuum can be expected (26). (2) A 45' angle of incidence was used to minimize overlap and hence interaction of the vaporization laser beam and the plume, which ejects in a direction perpendicular to the sample surface (27). Enhancement in line-to-continuum ratio has been reported with this geometry (28,29). (3)At pressures leas than 100Torr, the major part of plasma emission is localized near the crater. The line emission, on the other hand, emits from a larger volume (26,30). Thetefore, a spatial masking scheme was employed, by which the emission from the region near the crater was blocked by the inclined sample piece. Only the emission from the top periphery of the plume was collected. This improved to some extent the line-to-background ratio (31). (4) The most intense part of the continuum lasts for a relatively short time, while the line emission persists much longer (microseconds) after the decay of the coninuum emission (32.33). This is shown in Figure 2, which was ob-
tained by using a color glass filter instead of the monochromator to block the laser line. The spike with a time duration of about 1ps corresponds to the continuum plasma emission, whereas the emission a t the longer time period is from the line emission. Temporal discrimination (34,351 is therefore useful. This was also provided by the tilted sample itself, due to the time needed for the material blown off to travel from the surface to the top periphery to become observable. The time involved can be estimated to be in the microsecond regime. Time gating can be easily introduced in the waveform analyzer used, since it is very convenient to integrate the emission intensity over a selected time segment. (5) Since the spectral characteristics of the plasma are continuous, the use of a monochromator proved necessary to reject the unwanted radiation. (6) For optical breakdown of gases under an intense radiation field, the power threshold increases with the decrease in gas pressure (36),so it was essential to keep the buffer gas pressure low. Helium was used because of its relatively high breakdown threshold at reduced pressure. (7) Finally, it is essential to keep the power density low, since the ionization efficiency, governed by the Saha-Langmuir equation, is highly dependent on this parameter. Below lo8 W/cm2 the ion-to-neutral ratio is very low, of the order of lo6, while at 5 X 109 W/cmz, this ratio can approach 100% (23). In this work the power density was maintained to be less than lo9 W/cm2 as judged from the crater size. The inert gas introduced into the cell couples the acoustic wave to the microphone. As a filler gas, helium gives a higher acoustic signal than other gases at otherwise the same experimental conditions, due to its low heat capacity and high thermal conductivity (37). It had a second function here, to confine the plume to a small size so the masking scheme mentioned above could he implemented. In addition, inert gases have been found to enhance the emission signal and reduce the continuous hackground (38). Lastly, its chemical inertness avoided the complication from potential chemiluminescent reactions between the plume and the atmospheric gas (39). When the pressure is below 1Torr, the coupling efficiency is low and the acoustic signal is relatively weak and decays rapidly. The signal amplitude increased with pressure, due to the increase in coupling efficiency (40). A t a pressure of around IO Torr, it was found that the coupling efficiency increased sharply with laser power, resulting in an undesirable nonlinear response with respect to laser power. This is because the material evaporated (at high power) has a considerable contribution to the total pressure. The response tends to be linear at higher buffer pressures. But if the pressure approaches atmospheric pressure, complicated waveforms are observed. These result from rapid propagation, reflection, and mixing of the acoustic waves within the cell. So typical pressure within the cell was controlled at 50 Torr. The contribution of the sound wave transmitted through the solid media (sample mount and cell body, etc.) was found to be insignificant. Only minor peaks appeared in the time scale corresponding to the time needed to transmit the impulse to the microphone through the solid path. This was also confirmed by blocking the front path to the microphone with an aluminum disk. Therefore the acoustic wave observed had two pasible sources, a density wave and a thermal wave. The former originates from material evaporated by the pulsed laser, Le., an impulse generated by the explosive vaporization of ablated material from the solid to form the plume. The latter is due to heating of the solid and the buffer gas under laser radiation and is independent of the vaporization process. Experimentally,these two components could not be separated, since they reached the microphone a t about the same time.
‘-c::.
ANALYTICAL CHEMISTRY. V M . 60. NO. 20. OCTOBER 15. 1988 2201
-+ w
2c
4
Y
p.
u z
8 Q
0.4 0.2
o
0
4
8
I2
I6
20
24
28
EMISSION PEAK AREA (UVS)
Acwstic peak helght vs emlssion peak area at different fccuslng condnions. No time ganng was applied to emission slgnal. C”% sample. NBS SRM C115Oa whke case ton:laser. 308 nm. 1 Hr. 25 ns. 5 mJIpulse; buffergas,He at 50 Twr. The farthest poht at top right corresponds to exact focusing. Each adjacent point caresponds to 0 . 5 ” movement of the focusing lans away from the sample surface. Figure 4.
Waveformanalyzer screen display: upper trace, 403.1-nm Mn I emission at 40 nslpoint: lower trace. acoustic wave at 200 nslpoint. The start of the trace is synchronized with the laser pulse. Condeions: sample, NBS SRM C1288: laser. 308 nm, 25 ns. 1 Hz. 1.2 mJlpube: buffer gas. He at 50 Torr. Flgure 3.
However, as discussed below, there is evidence that the density wave predominates in the acoustic wave detected. Correlation of Signals. The correlation between acoustic and emission signals associated with the lasergenerated plume was investigated. These two signals were monitored simultaneously a t different laser powers, different focusing conditions, and different laser shots on various samples with different compositions or treated by different procedures. A good linear correlation will imply that the acoustic signal can be used to normalize emission signals to improve the precision in quantitative analysis. Detection of emission was chosen because of its simplicity (no additional excitation source was needed). The minor component Mn ( s c w
22 X
4
w
p.
u rl a 8 4
EMISSION PEAK AREA (uVS)
Fbum 5. Acoustlc peak helgM vs emlssion peak area at dlfferem focushg Conditions. The emission slgnal was Integrated over 1.6-20 ps. Omer conditions were aentical with hose in Figure 4.
Figure 4 shows that ACPH-EMPA correlation without time gating, while Figure 5 shows the ACPH-EMPA correlation in which the emission intensity was integrated only during the 1.6-20 w time period. Good Correlation (R2= 0.9922) was obtained in the second ease,whereas in the first case the data a t high power density revealed higher acoustic response and lower emission intensity than those expected from a linear function. This can be explained by a higher degree of plasma formation under high laser power density a t the very early stage (42). The absorption of laser energy by the plasma leads to attenuation of the laser beam reaching the sample surface (28)and the generation of heat and hence thermal expansion of the plume. The consequences are weaker emission and stronger acoustic signal. When the laser beam waist was moved away from the sample surface by 5 mm, both the emission and acoustic signals decreased by 2 orders of magnitude. Since the heating of the window, the buffer gas. and the solid matrix itself should not change much under the various focuaing conditions, one can conclude that the acoustic wave components associated with those processes did not contribute significantly to the total acoustic signal. The impulse from the material vaporized off the surface should then be the dominant mechanism in acoustic wave generation. Based on Figure 5, in the experiments that follow, the laser beam waist was placed 2 mm shove the sample surface to avoid excess plasma emission. The ACPH-EMPA correlation a t different laser powera (fixed focus) was also studied by using SRM C1150a white cast iron. Under experimental conditions similar to those in Figure 5, when the power was decreased gradually, both acoustic and emission signals were lowered by more than 2
ANALYTICAL CHEMISTRY, VOL. 60, NO. 20, OCTOBER 15, 1988
2262
/ 1
4001
'
" Y
a
I
I
4 0
I
2
4
6
8
EMISSION PEAK AREA (uVS)
Figure 6. Acoustic peak height vs emission peak area at different laser shots on the same spot on sample surface: sample, NBS SRM C115Oa white cast iron; laser, 308 nm, 1 Hr, 25 ns, 5 ml/pulse; buffer gas, He at 50 Torr. Data collection was performed at every shot at the beginning and then every 5-30 shots thereafter. A total of 430
shots were tracked. orders of magnitude, resulting in a good linear relationship with R2 = 0.9939. The limitations are microphone saturation at high power and background noise at low power. The latter came from sources such as the thermal fluctuation of the buffer gas, ambient acoustic turbulences including mechanical vibrations from vacuum pumps, window heating, and electronic noise associated with the preamplifiers and the detection system used (43). Linear correlation had been found for data extending over about 3 orders of magnitude but never beyond this range. Good correlation was found both for data obtained only from the fist laser shot (R2= 0.9948) and from several superimposed shots on the same spot on the sample surface (R2= 0.9939). These results imply that variations in emission due to laser power fluctuations can be properly corrected for by monitoring the acoustic signal. ACPH-EMPA correlation was investigated for an extended series of superimposed laser shots on the same spot on the surface. During repeated laser irradiation, the surface conditions of the irradiated spot change drastically. Heating causes vaporization and surface roughening. A crater with an annular rim was formed as a result of vaporization, melting, and redeposition (44).As has been observed previously (21), the signal from the same spot decreased after repeated exposures. This is shown in Figure 6. The amount of material removed in each laser shot can be as little as 1ng. So, data was collected at every shot a t the beginning of the series of vaporization events, with the collection interval gradually increased to every 30 shots thereafter. The correlation was tracked for several hundred laser shots, during which both signals decreased by more than 1order of magnitude. A fairly good correlation (R2 = 0.9710) was maintained. The correlation in Figure 6 is a bit worse (larger scatter from the line) than in the above experiments. As the depth of the crater increases, confinement of the ejected plume occurs (21). One would expect the acoustic wave to be similarly affected. This result, as well as those mentioned above, implies that one can normalize the emission signal by monitoring the associated acoustic wave when focusing and surface conditions are varied. This should be especially useful in depth profiling, where it is extremely difficult to keep sample movement exactly in pace with the increase in the crater depth. When the same measurements were made on two pieces of SRM C1288 high-alloy steel treated in different ways (sanded only, and further polished by using grade A and B polishing powders with 0.3- and 0.05-llm diameters, respectively), the two straight lines obtained for plots similar to Figure 6 showed similar slopes (0.0511 f 0.0014 vs 0.0469 f 0.0011 for 50 data points each). It was found that SRM C1222 Cr-Ni-Mo steel
and SRM C1150a white cast iron yielded slopes with similar values (0.0387 f 0.0014 vs 0.0399 f 0.0015 for 50 data points each), indicating both samples had similar Mn contents (0.78% vs 0.77% in this case), even though the physical properties and chemical compositions of these two samples are quite different. But, on SRM (21288 high-alloy steel (with a similar Mn content of 0.83%) a slope of 0.0469 was obtained under the same conditions. Since the reciprocal of the slope is a measure of the Mn content, this unusually high slope indicates spectral interference from the major component, Fe. SRM C1288 had an Fe content of 43%, whereas the others had Fe contents of more than 90%. This is a direct consequence of incomplete separation of the 403.1-nm Mn I triplet from adjacent Fe lines. As was mentioned in ref 41, the 403.1-nm Mn I line is immune to Fe interference at an analyte-to-concomitantratio of 1:200,so that difficulties with iron and steel analysis could be solved by using a monochromator with higher resolving power. These results show that variations in the amount of sample vaporized due to matrix differences and surface treatment can be corrected for by acoustic wave monitoring to a certain extent. Good correlation (R = 0.994) was also found between the acoustic wave and the emission from the major component. In this case, the Fe 386.0-nm line was used. This implies that the emission intensities of the major and the minor components are also well correlated. So, the use of intensity ratios for normalization (13,14) is a good method whenever it can be implemented. Obviously, one has to assume that the major (standard) element is uniformly distributed and that the two elements are excited equally. On the other hand, similar correlation could not be found between the acoustic signal and the continuous plasma emission. This is not surprising, since the plasma is highly absorbing and is excited by complex mechanisms. It is also interesting to note that very similar values for the slopes were obtained on different samples with similar Mn contents under widely varying conditions, even when red hot particles were observed during the experiment. These particles actually do not contribute to either emission or acoustic signals. Therefore, unlike the case of relying on crater size measurements, (12),good correlation could still be maintained here.
CONCLUSIONS Ultimately, it would be ideal to have a universal straight line by which all fluctuations in signal amplitude in lasergenerated plumes could be normalized, regardless of the experimental conditions. One must naturally exercise caution in using any one signal for normalization. We have shown here that the acoustic signal can track the emission signal linearly over 3 orders of magnitude. This technique will be especially valuable as an internal standard for normalization in depth profiling and in other LMA experiments in which the laser output, sample conditions, and other experimental parameters vary substantially. In this work only the emission signal was monitored, but application of this normalization procedure could be easily extended to other spectroscopic techniques like atomic absorption (45) and fluorescence (46) and to cases where the ablated material is swept to another location (47) or redeposited onto a collector for further excitation (48). LITERATURE CITED Laqua, K. In AnalyricalLaser Specboscopy;Omnetto, N., Ed.; Wiley:
New York. 1979; Chapter 2. Rasberry, S. D.; Scribner, B. F.; Margoshes, M. Appl. Opt. 1067, 6 , 87-93, Van Deijck, W.; Balke, J.; Maessen, F. J. M. J. Spechochlm. Acta, Pari 8 1070, 348, 359-369. Cerrai, E.; Trucco, R. Energ. Nucl. (Mlkn) 1068, 15, 581-587. Kirchheim, R.;Nagorny, U.; Maimr, K.; Tolg, G. Anal. Chem. 1076, 48, 1505-1508. Karn. F. S.; Singer, J. M. Fuel 1068, 47, 235-240.
Anal. Chem. 1088, 6 0 , 2263-2268 Wennrich, R.; Dktrich, K.; Bonk, U. Spectrochlm. Acta, Part 8 1984, 398, 657-666. Dktrich. K.; Wennrich, R. Prog. Anal. At. Spectrosc. 1984, 7 ,
139-198.
Felske, A.; Hagenah, W.D.; Laqua, K. Spectrochim. Acta, Part B 1972, 278, 1-21. Pleameler. E. H. I n AnaMicel ADDlicetions of Lasers: PieDmeier. E. H., Ed.: wileyi -New York. 7988;C k p t e r 19. I Peppers, N. A,; Scribner. E. J.; Aiterton, L. E.; Honey, R. C.; Beatrice, E. S.; Harding-Barlow, 1.; Rosan, R. C.; Glick. D. Anal. Chem. 1988, 40, 1178-1182. (12) Morton, K. L.; Nohe, J. D.; Madsen, B. S. Appl. Spectrosc. 1973, 27,
109-117. (13) Carr, J. W.; Horlick, G. Spectrochlm. Acta, Part 8 1982, 378,1-15. (14) Talmi. Y.; Sieper, H. P.; Moenke-Bankenburg, L. Anal. Chlm. Acta 1961, 127. 71-85. (15) Saffir, A. J.; Marlch, K. W.; Orenberg, J. B.; Treyti, W. J. Appl. Spectrosc. 1972, 2 6 , 489-471. (16) . . HOuk. R. S. I n AnaMlcel Appllcatlons of Lasers; Piepmeler, E. H., Ed.; Wiley: New York, 1986;Chapter 18. (17) Allemand. C. D. Spectrochlm. Acta, Part8 1972, 278, 185-204. (18) Margoshes, M.; Marcellus, D. A,; Rasberry. S. D. froceedlngs of the 13th Colloquium Spectroscoplcum Internatlonale; Hilger:
London,
ln68: D 156. (19) B&u&v, V. I. Zh. Anal. Khim. 1984. 3 9 , 909-927. (20) Bykovskii, Yu. A.; Basova, T. A.; Belousov, V. I.; Giadskol, V. M.;
Qorshkov, V. V.; Degtyarev, V. G.; Laptev, I.D.; Nevolin, V. N. Sov. Phys. Tech. Phys. (Engl. Trans/.) 1978, 2 1 , 761-763. (21) Chen, G.; Yeung. E. S. Anal. Chem. 1988, 60. 864-868. (22) Carroll, P. K.; Kennedy, E. T. Contemp. Phys. 1981, 22, 61-96. (23) Conzemius, R. J.; Capellen, J. M. Int. J. Mass Spectrom. Ion Phys. 1980, 3 4 , 197-271. (24) Fabbro, R.; Fabre, E.; Amiranoff, F.; Garban-Labaune, C.; Virmont, J.; Weinfeld, M.; Max, C. E. Phys Rev. A 1982, 2 6 , 2289-2292. (25) Browne, P. F. R o c . Phys. SOC. 1985. 8 6 , 1323-1332. (26) Piepmeler, E. H.; Osten, D. E. Appl. Spectrosc. 1971, 2 5 , 642-652. (27) Knox, B. E. I n Trace Analysls by Mass Spectrometry; Ahearn, A. J., Ed.; Academic: New York. 1972;Chapter 14. (28) Plepmeier, E. H.; Malmstadt, H. V. Anal. Chem. 1969, 4 1 , 700-707. (29) Dimitrov. G.; Zheleva, T. Spectrochim. Acta, Part B 1984. 398,
.
.-- - .- .- .
1209-1219
(30) Sucov, E. W.; Pack, J. L.; Pheips, A. V.; Engelhardt, A. G. Phys. Fluids 1987, 10, 2035-2048.
2263
(31) Treytl, W. J.; Marich, K. W.; Gllck, D. Anal. Chem. 1975, 47, 1275-1279. (32) Archbold, E.; Harper, D. W.; Hughes, T. P. 8 r . J. Appl. Phys. 1984, 15, 1321-1326. (33) Lltvak, M. M. Edwards, D. F. I€€€ J. Ouantum Electron. 1988, E - 2 , 486-492. (34)Scott. R. H.; Strashelm, A. Smctrmhlm. Acta, Part 8 1970, 2 5 8 , 311-332. (35) Treyti, W. J.; Orenberg, J. B.; Marich, K. W.; Glick, D. Appl. Spectrosc. 1971, 25. 376-378. (36) Minck, R. W. J. Appl. Phys. 1984, 35, 252-254. (37) Adams, M. J.; King, A. A,; Kirkbright, G. F. Analyst (London) 1978, 101.73-85. (38) Treyti, W. J.; Marich, K. W.; Orenberg, J. 8.; Carr, P. W.; Mllier, D. C.; Glick, D. Anal. Chem. 1971, 43, 1452-1456. (39) Beenen, G. J.; Piepmeier, E. H. Appl. Spectrosc. 1984, 3 8 , 851-857. (40) Antlpov, A. B.; Kapitanov, V. A,; Nikiforova. 0. Yu.; Ponomarev. Yu. N.; Sapozhnlkova, V. A. J. Photoacoust. 1983-1984, 1 , 429-443. (41) Kriege, 0.H.; Marks, J. Y.; Welcher, 0. 0. I n Flame Emission and Atomic Absorption Spectroscopy, Vol. 3 , Elements and Matrlces ; Dean, J. A., Rains, T. C., Eds.; Dekker: New York. 1975;Chapter 7. (42) Conzemius, R. J.; Zhao, S.; Houk, R. S.; Svec, H. J. Int. J. Mass Spectrom. Ion Processes 1984, 6 1 , 277-292. (43)Zharov, V. P.; Letokhov, V. S. I n Laser Ontoacoustlc Spectroscopy; Tamir, T., Ed.; Springer Series in Optical 7 :iences, vol. 37;SpringerVerlag: New York, 1984;Chapter 6. (44) Wagner, R. E. J. Appl. Phys. 1974, 45, 4631-4637. (45) Manabe, R. M.; Plepmeler, E. H. Anal. Chem. 1979, 51, 2066-2070. (46) Measures, R. M.; Kwong, H. S. Appl. Opt. 1979, 18, 281-286. (47) Kahtor, T.; Bezur, L.; Pungor, E.; Fodor, P.; Nagy-Balcgh, J.; Heincz, 0. Spectrochlm. Acta, Part 8 1879, 348, 341-357. (48)Rudnevsky, N. K.; Tumanova, A. N.; Maximova, E. V. Spectrochlm. Acta, Part B 1984, 398, 5-11.
RECEIVED for review May 20,1988. Accepted July 13, 1988. The Ames Laboratory is operated by Iowa State University for the U.S.Department of Energy under Contract W-7405Eng-82. This work was supported by the Director of Energy Research, Office of Basic Energy Sciences, Division of Chemical Sciences.
Catalytic Reduction of Myoglobin and Hemoglobin at Chemically Modified Electrodes Containing Methylene Blue Jiannong Ye and Richard P. Baldwin* Department of Chemistry, University of Louisville, Louisville, Kentucky 40292
Chemically modified electrodes (CMEs) exhibiting eiectrocatalytic response toward myoglobin and hemogiobln were constructed by adsorbing the phenothiazine mediator titrants methylene blue and thionine onto spectroscopic graphite. These CMEs, which were prepared by a rapid (60 s) and reproducible (3.2 % relative standard deviatlon) dip-coating procedure, permilted the hemoprotein electroreduction to take place at the reduction potential of the mediator molecule. For neutral and slightly acidic solutions, this corresponded to very modest negative potentials (vs Ag/AgCi). When used in flow injectkm and iiquld chromatography detection, with an applied potential of -0.12 V, the methylene blue CMEs gave detection ilmits of 10 and 20 pmol Injected for myoglobin and hemoglobin, respectively, with linear response extending 2 to 3 orders of magnitude higher. After a brlef equilibration period, the CME retained more than 90% of its inltlai myoglobin response over several hours of contlnuous exposure to the chromatographic flowstream.
Although many enzymes and proteins possess functional 0003-2700/86/0360-2263$01.50/0
groups that can be readily oxidized or reduced by chemical redox agents, it is rare for these same compounds to undergo facile oxidation or reduction at electrodes. Rather, for reasons ascribed either to their extended three-dimensional structure and the resulting inaccessibility of the eledroactive center or to their adsorption onto and subsequent passivation of the electrode surface, most biological macromolecules exhibit such slow rates of electron transfer that no useful currents are observed at conventional electrodes, even with the application of relatively large overpotentials. To date, only a few exceptions to this behavior have been demonstrated, with these occurring primarily when rather specialized electrode materials were employed. For example, reversible or quasi-reversible electrolysis of cytochrome c has been reported at indium oxide (1-3), tin oxide (3), and edge-graphite (4), and the use of ruthenium oxide has been shown to facilitate the electrode reactions of several proteins including cytochrome c , azurin, rubredoxin, ferredoxin, and plastocyanin (5). However, even in the most favorable of these cases, the observed electrochemistry was strongly dependent on pH, electrolyte, and other solution conditions. In a few instances, bare gold ( 3 ) , platinum ( 3 ) ,and silver (6) have been reported to give rela0 1988 American Chemical Soclety
