1. Introduction

1.1 Technological background

Many high temperature processes of technological importance rely upon, or are adversely affected by, chemically active heterogeneous subsystems. One such case involves chemical interaction of a multicomponent high temperature (e.g., ~500-2000 K) high pressure (e.g., ~0.1-100 atm) gas mixture with a solid or liquid substrate. This interaction frequently leads to the formation of intermediate species containing elements of both the gaseous and substrate materials. These reaction intermediates can be transported in the presence of temperature, concentration, or momentum (e.g., forced convection) gradients. Transport to another regime of temperature, pressure, or concentration can lead to a reaction-reversal or the introduction of secondary processes. In many instances, this changing chemistry along gradients results in a deposition of intermediate species with the overall result that the initial substrate has been physically transported to another part of the system.

Such material transport can have either an undesirable or beneficial effect on the system of interest. Examples in modern technology include, respectively:

(1) Hot corrosion of gas turbines, jet engines, rockets, coal gasifiers, magnetohydrodynamic channels, coal fired boilers, and numerous pyrometallurgical systems. (2) Controlled material transport for production of crystals and films, extractive metallurgy, flame inhibition and fire extinguishment, combustion modification such as smoke and antiknock control with additives, and regenerative lamp cycles.

A detailed recent discussion of these and other examples of high temperature materials transport has been given elsewhere [1]1.

In order to understand these heterogeneous processes for development of new or improved control strategies, it is necessary to define at a molecular level the transport mechanisms and particularly the reaction intermediates. To date, extensive reliance has been placed on speculation and empirical observations based solely on macroscopic variables defining the initial and final state of the system. For instance, current understanding of hot corrosion derives from metallurgical examination of the clean and corroded material under ambient conditions. Extrapolation of these observations to the actual conditions of usage requires many assumptions about the corrosion mechanisms. However, detailed mechanistic information is practically nonexistent. A further inaccuracy occurs because standard test procedures (e.g., use of laboratory burner rigs) simulate only part of the actual system. To properly relate test results to practical usage would require a knowledge of scale-up factors derived from a detailed mechanistic description of both the test and the end-use systems.

Information on reaction mechanisms, including the intermediate transport species, has been practically nonexistent heretofore because of the measurement problems associated with these extreme conditions. Very limited molecular-specific information can be obtained using optical spectroscopic methods at high temperature and pressure. In principle the broad spectrum of gaseous and vaporized species present could be identified and their spatial and temporal concentrations determined by the high pressure sampling mass spectrometric (HPMS)

1Figures in brackets indicate the literature references at the end of this paper.

technique. This molecular beam technique, which uses a sonic nozzle for sample extraction, has been amply demonstrated by us [2] and others for homogeneous systems, including flames. A summary of recent work in this area has been given by Stearns, et al. elsewhere in this volume [3]. For heterogeneous systems, the gas is saturated by condensible species. In this case, the nozzle probes used to interface the hot gas and the mass spectrometer become part of the reaction system and severe limitations arise. Conventional use of the HPMS technique (homogeneous systems) has required probe temperatures to be considerably less than for the sample system. With saturated gases, this leads to condensation of inorganic species at the probe tip and results in physical blockage of the small entrance orifice or in corrosive loss of the probe.

We have developed a procedure for avoiding this difficulty, in addition to providing other advantages, as described in Section 2. Basically, the sample, probe and skimmer (to a lesser degree) are maintained isothermal by an external heat source. Maintenance of a steady state (and constant pressure) between gas and substrate requires a continuous replenishment of the gas extracted by the sampling process. This leads naturally to a transpiration procedure, where the input flow rate matches the gas extraction rate through the small-orifice probe. The technique therefore combines the basic features of both the transpiration [4] and the HPMS methods.

1.2 Scope of application

In the following sections we describe this technique of transpiration mass spectrometry (TMS) and its application to the characterization of high temperature vapors. We have demonstrated that it is a nonperturbing method2, which is always an important concern with probe

The method extends the dynamic range of classical vaporization and vapor transport techniques by many orders of magnitude; at least four orders with respect to Knudsen effusion mass spectrometry. We believe that TMS should be applicable as a quantitative measurement tool for laboratory simulations of most of the technological transport systems mentioned above and eventually may be applicable to plant-scale systems.

With regard to systems of initial academic interest, one can most likely expect this technique to be a useful means of producing novel high temperature species in a spectroscopically cool form, resulting from free jet expansion cooling. This would greatly extend the utility of thermally sensitive molecular beam spectroscopic methods. In particular, electron diffraction, microwave, laser fluorescence, Raman, photoelectron, and photoionization spectroscopy, as well as matrix isolation methods, would benefit from a coupling with the TMS technique. The extended high pressure range should also allow for the study of basic thermodynamic properties of novel complexes and adducts of the type suggested indirectly from gas-solid solubility studies (e.g., see Hastie, pp. 126-148; pp. 73-87 [1]). Future extension to even greater pressures should permit definitive molecular characterization of systems in their critical state and greatly supplement equation of state studies and the fundamental understanding of fluids.

2At least for the equilibrium systems studied, and this is most likely the case for many nonequilibrium systems.

2. Apparatus

2.1 The mass spectrometer system

Figure 1 shows a schematic of the mass spectrometer (stage II) and vacuum system layout with the transpiration apparatus attached (stage I). The aluminum walled vacuum system is partitioned into two differentially pumped stages. The stage labeled I utilizes a 6 in diffusion pump (Varian/NRC VHS-6)3 with a cold trap (Granville-Phillips Model 270-6) and gate valve (NRC 1279-6) close-coupled to the base of the vacuum chamber. Net pumping speed at the bottom of the chamber is calculated to be ~800 1/s. The backing system for this pump-stack uses a Roots blower (Leybold- Hereaus WA250) and a rotary two stage vane pump (Sargent-Welch 1375). Use of a blower in series maintains, even for high gas loads, a pressure at the

-4

diffusion pump outlet well below the ~104 atm1 critical backing pressure. This assures

-5

that the diffusion pump will continue to pump near its rated speed into the 10 atm pressure range. A high sensitivity thermocouple gage controller (Hastings-Raydist model SL-1R) and tube (type DV-8M) was used to monitor this pressure range.

3 Commercial sources are identified only for purposes of accuracy and do not imply endorsement by the U. S. Government nor that they are the most suitable product for the work performed.

41 atm = 1.01325 × 102 kpascal; see introduction to this volume for SI unit conversions.

A mechanical shutter separates stages I and II, acting as a valve to isolate the stages during removal of the transpiration reactor assembly. In addition, there are a variety of differential pump apertures (three hole sizes) in the shutter plate for control of the molecular beam diameter and stage I to II pressure difference. The aperture used for transpiration experiments is 0.05 cm in diameter and is located 4 cm from the skimmer orifice.

Stage II, the high vacuum mass spectrometer section, is pumped by a stack consisting of a 4 in diffusion pump (Varian/NRC VHS-4), a 4 in cold trap (Granville-Phillips Model 270-4) and a butterfly valve (Vacoa BFV-4). This stage has an approximate pumping speed of 400 1/s

at the vacuum chamber base. No-load vacuum in both stages is <5x10-11 atm. The background mass spectrum at this pressure shows the 18 amu (H20*) ion as the major peak (~50 percent of total ions) which arises from H20 desorption from the aluminum-walled vacuum system. At maximum load (~50 sccm [standard cubic centimeters per minute]) the stage I and II pressures -9 are ~5x10-6 atm and ~5x109 atm, respectively. A titanium sublimation pump (GranvillePhillips TSP series 287 in a 214-669 chamber) and an ion pump (Ultek Model 10-250), not shown in figure 1, are used as holding pumps on the stage II section when the diffusion pump is valved off to cycle the cold trap. This periodic trap cleaning procedure minimizes release of trapped reactive gas back into the mass spectrometer chamber.

The beam modulation chopper system, located in the stage II region, is a toothed wheel (3 or 24 teeth) driven by an AC synchronous motor (Globe 75-121-2) which has been cleaned for vacuum service. A reference signal, needed for phase sensitive detection, is generated by the wheel interrupting light from a small incandescent bulb (GE number 40) impinging on a photo transistor (GE-L14B). A sine wave oscillator (Donner model 1202) supplies the drive frequency to a chopper drive system [Extranuclear Laboratories (EL)] which produces the capacitively shifted two phase 115 V power to drive the chopper motor. The three tooth chopping wheel yields beam modulation frequencies from 40 to 270 Hz. Higher frequencies can be obtained with the twenty-four tooth wheel. The actual chopping frequency is normally selected by observing a very weak modulated beam mass spectral peak and adjusting for minimum noise interference from the 60 Hz line frequency. For these studies a chopping frequency of 2141 Hz was used.

The mass spectrometer itself is a quadrupole mass filter (EL-Model 270-9) with a crossedbeam configuration ion source (EL Model 041-2). We have modified the commercial ion source. by mounting a stainless steel plate at the ion source entrance for additional neutral beam collimation. Standard quadrupole power supply (EL model 011-1) and ionizer control electronics (EL type II) are used. The mass analyzed ion beam is deflected to a channeltron electron multiplier (Galileo Electro-Optics, model 4700) mounted off axis and operated at a nominal 1700 volts dc. Since the final amplifier is a frequency and phase sensitive lock-in detector (Ithaco Model 393), the multiplier gain is set for optimum signal to noise ratio at the lock-in, rather than for maximum gain. Total system gain using a fast electrometer amplifier (EL Model 031-2) is ~4x1011; signal to noise, rather than total signal being the sensitivity limiting factor.

This mass filter system is supported by a 6 in flange (conflat) mounted on a large welded bellows assembly which allows three-axis positioning of the ion source entrance

aperture. In practice, a pinhole and collimator alinement assembly is mounted on the beam axis with a He-Ne laser directed along the same axis. The mass filter assembly is then positioned so that the ion source entrance and exit apertures are centered on the laser beam. Final molecular beam alinement is achieved by adjusting both the mass spectrometer and the molecular beam source angle position to obtain maximum ion intensity of a non-scatterable beam component. This alinement procedure proved to be a critical phase of the experiment.

2.2 The transpiration reactor

The transpiration reactor is located in stage I of the overall assembly (see fig. 1). Essential features of this reactor include: a sample container or boat, a boat carrier, a thermocouple for temperature measurements, a carrier gas inlet system, and a gas extraction system or probe. The boat carrier allows for boat removal from the reactor without need for a complete disassembling of the transpiration system. Molecular beam sonic probes are typically conical nozzles with design details determined by reasonably well established gas dynamic criteria, as outlined in Section 3. However, for highly reactive systems, we also found it desirable to develop a relatively more robust capillary probe, at the possible expense of sampling fidelity. Pertinent gas dynamic considerations for this type of probe are given in Section 2.3. Below we describe the construction details of these probes and the transpiration assembly in general.

Figure 2 shows a schematic view of a typical transpiration reactor assembly, including details of the assembled boat carrier and boat. A platinum tube, with outside diameter 1.23 cm, wall thickness 0.075 cm, and length 5 cm is used as a boat carrier. This tube has machined ridges at each end to allow a snug fit with the main chamber walls for prevention of diffusion losses. Welded to the rear of the carrier is a section of 0.64 cm diameter platinum tube through which the carrier gas passes. This carrier gas tube is sealed to an SS304 1/4 in stainless steel tube via a SWAGELOK joint machined to fit inside the 1.27 cm i.d. of the main chamber. The seal is 21 cm from the front of the boat carrier. This long path length allows the gas to reach thermal equilibrium with the walls before interacting with the sample. The platinum section of the carrier gas tube also serves as a guide for introduction of an alumina sheathed Pt-Pt/10 percent Rh thermocouple to the sample area.

The boat, which is fabricated from welded platinum sheet, has a slanted base (~5° to the front) to allow liquid samples to be preferentially retained in the front of the boat (fig. 3). The gas baffle shown in figure 2 is located ~0.75 cm behind the boat, and serves to interrupt the fast gas flow into the boat region, thereby allowing the gas maximum time to equilibrate thermally before passage at a relatively slow rate over the sample. baffle also supports the tip of the alumina thermocouple sheath so that only bare wires are brought into the boat region, thereby minimizing exposure of potentially reactive ceramic to the system.