343
Anal. Chem. 1983, 55, 343-348
accurate mass measurement capabilities could be extended to higher masses. In future work, we intend to test this premise by carrying out similar studies for molecules with masses above 1000 amu under higher magnetic field conditions.
6-11, 1982, Honolulu, HI, paper WOC2. (4) Ledford, E. B., Jr.; White, R. L.; Kulkarnl, P. S.; Spencer, R. B.; Ghaderi, S.; Wilkins, C, L.; Gross, M. L. Abstracts of the 31st Pittsburgh Conference on Analytical and Applied Spectroscopy, March 5-9, 1979, Cleveland, OH, p 117. (5) Gross, M. L; White, R. L; Ghaderi, S.; Ledford, E. B„ Jr.; Wilkins, C. L. Abstracts of the 32nd Pittsburgh Conference on Analytical and Applied Spectroscopy, March 10-14, 1980, Atlantic City, NJ, p 462. (6) Harrison, A. G.; Jones, E. G.; Gupta, S. K.; Nagy, G. P. Can. J. Chem.
ACKNOWLEDGMENT
1966, 44, 1967-1973. (7) Scheffrahn, R. H.; Gaston, L. K.; Sims, J. J.; Rust, . K., submitted to
We wish to thank R. H. Scheffrahn, . K. Rust, and T. H. Morton for supplying real analysis problems requiring exact mass measurements.
J. Chem. Ecot.
Received for review August 23,1982. Accepted November 8, 1982. The support of the National Science Foundation under Grants CHE-79-10263, CHE-81-13612, and CHE-8018245 is gratefully acknowledged. This work was also supported in part by the Public Health Service under Grant
LITERATURE CITED (1) Comlsarow, . B.; Marshall, A. G. J. Chem. Phys. 1976, 64, 110. (2) Ledford, E. B,, Jr.; Ghaderi, Sahba; White, R. L.; Spencer, R. B.; Kulkarnl, P. S.; Wilkins, C. L; Gross, M. L. Anal. Chem. 1980, 52,
463-468.
(3) Allemann, M.; Kellerhals, Hp.; Wanczek, K. P. Proceedings of the 30th Annual Conference on Mass Spectrometry and Allied Topics, June
GM-30604.
Photoacoustic Cell for Fourier Transform Infrared Spectrometry of Surface Species John B. Kinney1 and Ralph H. Staley*1 Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
In particular, essentially no sample preparation is necessary. The properties of PAS make it a potentially valuable tool for the study of surface species. This capability has been demonstrated in both the ultraviolet-visible-near-infrared (7-9) and mid-infrared (10,11) regions by using conventional PAS. In this paper, we demonstrate the utility of FTP AS for obtaining mid-infrared absorption spectra of surface species. Even with the signal-to-noise enhancement of Fourier transform techniques, study of flat surfaces is not practicable. It is essential to have a sample with a moderate surface area, >1 m2/g, with higher surface areas desirable. Powder samples generally provide adequate surface areas. The powders used in these experiments generally have high surface areas, 10-100 m2/g. Examples are given of spectra of surface species on two types of surfaces, oxide powders and supported-metal powders. The oxide powder used here is a high-surface-area silica powder. The supported-metal powders are commercially available catalysts with a loading of about 5% metal on the alumina. It is useful to understand the relationship between the photoacoustic signal strength for surface absorptions and the amount of light absorbed. In conventional transmission spectrometry, Beer’s law states that A abc, where A is absorbance, a is absorptivity, b is the path length through the sample, and c is concentration. For a solid sample this may be rewritten as A = 06 where ß is the optical absorption coefficient. Theoretical analysis of the photoacoustic signal generated by a one-dimensional system shows that the signal intensity depends on a number of parameters in a complicated expression (12-14). Fortunately, it is often possible to make approximations that greatly simplify the final expression. Three parameters that are important in making approximations in the theoretical expression are ß, b, and µß, the thermal diffusion length of the sample. For a monolayer of molecules on a surface, only a small fraction of the light is absorbed in a single pass through the surface layer, indicating that the
A photoacoustic cell designed for chemical studies Is described which Incorporates temperature and atmosphere control. The design utilizes a Helmholtz resonance to en-
hance sensitivity over a broad acoustic frequency range for use with Fourier transform Infrared (FTIR) Instrumentation. The effective efficiency at resonance Is 90 Pa/W. Fourier transform Infrared spectrometry with photoacoustic detection Is demonstrated as a useful technique for obtaining mid-infrared spectra of surface species. Examples are given of molecules on silica powder and supported metal catalysts. The quantitative dependence of signal Intensity on various factors Is discussed.
Photoacoustic spectroscopy (PAS) has, in recent years, been developed into a useful technique for the study of the optical and thermal properties of solids (1,2). Most of the work done with PAS has been in the ultraviolet-visible region of the spectrum. This is primarily because PAS requires a moderately intense light source. Extension of PÁS into the midinfrared region has been difficult due to the absence of widely tunable infrared light sources with sufficient power to make PAS a practical technique for obtaining mid-infrared spectra. Fourier transform infrared (FTIR) spectrometry with photoacoustic detection has recently been demonstrated as a useful technique for obtaining mid-infrared spectra (4000-500 cm-1) of solids (3-6). The advantages inherent in Fourier transform techniques more than compensate for the low light intensity of the glow-bar source used in these instruments. The combination of FTIR with PAS makes it possible to obtain infrared spectra of solids with the convenience of PAS.
=
Current address: Central Research and Development, E. I. du Pont de Nemours and Co., Experimental Station, E356, Wilmington, DE 19898. 1
0003-2700/83/0355-0343S01.50/0
©
1983 American Chemical Society
344
·
ANALYTICAL CHEMISTRY, VOL. 55, NO. 2, FEBRUARY 1983
surface layer is “optically thin”. The thermal diffusion length, µ8, is much larger than the thickness of a monolayer, so the sample is also “thermally thin”. In these limits, the expression
for the signal intensity simplifies to
s
=*
ß
-rY °g°b"b
(1)
where og and ob are the thermal diffusion coefficient of the gas and support, respectively, kb is the thermal conductivity of the support, and Y is an instrumental constant (12). The thermal diffusion coefficient is proportional to the square root of the modulation frequency, 1/2, so the signal is proportional to "1. This expression was derived with a simple one-dimensional model and is not necessarily valid for a powdered sample. This expression has been tested by using samples of absorbers supported on high surface area powders. A cell for photoacoustic (PA) detection is conceptually simple. All that is necessary is an acoustically sealed chamber that contains a window, the sample, and a microphone. However, various factors must be considered in order to design an efficient and useful cell. These factors fall into three categories: those that maximize the signal intensity, those that minimize the background signal and noise, and those that determine convenience and versatility. There are several factors that affect the signal intensity. The most important factor is the internal cell volume. Signal intensity is inversely proportional to cell volume so it is desirable to have the smallest volume possible within the limits set by other factors (15). Another important factor is the distance between the window and the sample. If this distance is less than the thermal diffusion length of the gas, heat generated at the sample will be lost to the window, decreasing the signal intensity (12). Separation of the window and sample by a distance of 2 mm is sufficient to prevent this effect with most gases for modulation frequencies above 10 Hz. Signal intensity can also be lost due to acoustic damping at small channels in the cell (16,17). This damping results from viscous friction in the gas and heat losses to the walls during sound transmission. This is usually only a problem when the microphone is at the end of a small diameter tube as in resonant cell designs. Minimizing the background signal and noise is an important consideration in PA cell design. Two factors are important for minimizing the background signal. First, scattered light must be kept from striking the microphone diaphragm since the microphone is extremely sensitive to signals generated by modulated light absorbed by the diaphragm (18). Second, light absorption by the interior surfaces of the cell must be minimized. PAS cells are, therefore, usually made of either polished metal or optical-grade glass. There are two important sources of noise in a PA cell: mechanical vibrations and acoustic leaks. The microphone diaphragm is particularly sensitive to vibrations. This effect is minimized by mounting the entire cell assembly on a base plate that is vibrationally isolated from the table. The acoustic leaks in the cell must also be minimized. These leaks not only increase the background noise but also decrease the signal intensity. One acoustic leak that is often overlooked is the back of the microphone. Back-vented microphones have a small diameter tube between the back of the diaphragm and the preamplifier that keeps the region behind the diaphragm at ambient pressure. While this channel is highly damped at audio frequencies, high quality microphones can easily detect room noise through this channel. A back seal that increases the damping of this noise source can greatly improve the PA cell performance. Even front-vented microphones have acoustic leaks around the rear connectors and must have the damping increased. If these precautions are taken, the noise level
should be just that due to the preamplifier noise. A PA cell that is to be used in the study of reactive chemical systems must incorporate several features. First, it is necessary to be able to control the gas atmosphere in the cell. This is often required to protect the sample from reaction with oxygen or water. Second, it is useful to be able to load the sample into the cell in a controlled atmosphere glovebox. Third, it is convenient to be able to use the sample cell itself as the reaction vessel for the preparation and reaction of a sample by controlling both the pressure and temperature of the gas in contact with the sample. A cell with an acoustic resonance can be useful for enhancing the sensitivity of the cell (19, 20). The most useful resonant design for a PA cell is a Helmholtz resonator (21). This design consists of two chambers connected by a long narrow tube. Helmholtz resonances in PA cells have been discussed by several authors (22-25). The resonant frequency can be determined by the expression
Fr
=
rC 2ttV2
Vi + ^2
V^L
1/2
+ 1.7 r)
(2)
where C is the speed if sound, Vj and V2 are the volumes of the two chambers, and L and r are the length and radius of the connecting tube (21). This type of resonator has the advantage that a moderately low resonance frequency can be achieved while maintaining a low internal volume. Many different designs have been proposed in the literature for use as PA detection cells (2). These designs range from simple multipurpose cells (2,14, 26-29) to elaborate special purpose designs (20, 30, 31). However, none of these designs in the literature are suitable for use in studies of reactive chemical systems. The PA cell described in this paper incorporates all of the features discussed above for studies of this kind while still achieving high sensitivity and low noise.
EXPERIMENTAL
SECTION
The spectra presented here were obtained on a Nicolet 7199A FTIR spectrometer using the photoacoustic cell described below. In FTIRPAS mirror velocity establishes a correspondence between wavenumber and acoustic frequency. A slow mirror velocity was used, typically 0.068 cm/s, so that the wavenumber range (600-4000 cm"1) corresponds to the acoustic frequency range 80-540 Hz. This cell allows control of the composition of the gas in the cell and of the temperature of the sample in the range -60 °C to +100 °C. The sample chamber can also be used as a reaction vessel. Highly reactive, air-sensitive samples are easily handled. The gases used in these experiments were supplied by Matheson and were used without further purification. The inorganic support, silica (400 m2/g), from Alfa products was heated to 400 °C under vacuum before use. The supported metal catalysts, 5% metal on alumina, supplied by Alfa and Strem, were cleaned by cycling under hydrogen and oxygen at 200 °C. The silane reagents were supplied by Petrarch. The organic solvents were dried by refluxing over P205 before distilling. Photoacoustic Cell Design. A cross section of the PA cell is shown in Figure 1. The basic design consists of two small chambers connected by a tube. The volumes of the sample and microphone chambers are kept as small as possible, ~0.25 cm3 each. With the volume of the connecting tube, the total volume is ~0.75 cm3. Typically, the cell is loaded with a sample layer 1.0 mm thick. This results in a window-sample spacing of 1.6 mm, a practical compromise between improved low-frequency response obtained with larger window-sample spacing and improved overall response obtained with lower volume with smaller window-sample spacing. The removable sample cup is loaded into the sample block from the top, with the window sealed over it with an O-ring. The gas atmosphere can be controlled by attaching a gas line to the cell. Valves are incorporated into the two tubes leading from the sample chamber to the gas line and to the microphone chamber. The sealed sample chamber can be detached from the microphone chamber and can readily be attached to a vacuum line or taken into a glovebox. Temperature control is achieved
ANALYTICAL CHEMISTRY, VOL. 55, NO. 2, FEBRUARY 1983
Figure 1. Cross-section of the PAS cell designed in this laboratory: [A] sample block; [B] removable sample holder; [0] cover plate; [D] 1-in. diameter window, NaCI or ZnSe for mid-IR, sapphire or Suprasil for UV-vislble; [E] microphone block; [F] B&K 4166 1/2 in. condenser microphone; [G] B&K 2916 preamplifier; [H] preamplifier acoustic seal; [I] Viton O-rlngs; [J] Hamilton miniature Inert valves; [K] Hamilton Luer-Lock connectors.
·
345
.5
Figure 3. Signal-to-noise ratio vs. frequency. The noise level increases by a factor of 2 at the resonant frequency indicating that the cell is not acoustically isolated.
Table I. Photoacoustic Signal Intensity and Frequency Response with Various Gases gas
c3h6
N,
Ar He h2
Mg,a mm
0.14 0.29 0.32 0.45 0.76
s/ 7Mgb
C,c m/s
Fx, Hz
Fx/Cd
34 22 18 12
250 334 319 965 1284
425 550 475 1225 1675
1.7 1.6 1.5 1.3 1.3
7
Figure 2. Frequency response of PAS cell: (·) cell filled with dry nitrogen (The straight line Is the expected response of a nonresonant cell having the same volume S
