Scholarly article on topic 'Safety-related conclusions for the application of ultrasound in explosive atmospheres'

Safety-related conclusions for the application of ultrasound in explosive atmospheres Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Keywords
{Ignition / Explosion / Cavitation / "Standing wave fields" / "High intensity ultrasound"}

Abstract of research paper on Chemical engineering, author of scientific article — Lars Hendrik Simon, Volker Wilkens, Michael Beyer

Abstract International safety regulations such as EN 1127-1 consider ultrasound to be an ignition source. Currently, applications of ultrasound in explosive atmospheres have to comply with a threshold value of 1 mW/mm2. However, it is unclear as to how this intensity has to be measured and, therefore, this threshold value is poorly defined. Moreover, it is based on theoretical estimations in analogy to other ignition sources and there are no publications or significant records on these estimations. Within a research project at PTB, it has now been investigated experimentally in relation to worst-case considerations including airborne ultrasound, focused MHz ultrasound in liquids and acoustic cavitation. On the basis of the results of the research it is now possible to revise the current regulations and to specify measures for safe operation of ultrasonic applications in explosive atmospheres. In this context, for ultrasound coupled directly to gaseous atmospheres a new threshold value of 170 dB (re. 20 μPa) can be suggested, and for ultrasonic applications in liquids, an augmentation can be made to the threshold to 400 mW/mm2.

Academic research paper on topic "Safety-related conclusions for the application of ultrasound in explosive atmospheres"

Contents lists available at ScienceDirect

Journal of Loss Prevention in the Process Industries

journal homepage: www.elsevier.com/locate/jlp

Safety-related conclusions for the application of ultrasound in explosive atmospheres

Lars Hendrik Simon, Volker Wilkens, Michael Beyer*

Physikalisch Technische Bundesanstalt (PTB), Bundesallee 100, 38116 Braunschweig, Germany

CrossMark

ARTICLE INFO

Article history:

Received 15 September 2014 Received in revised form 8 March 2015 Accepted 9 March 2015 Available online 11 March 2015

Keywords: Ignition Explosion Cavitation

Standing wave fields High intensity ultrasound

ABSTRACT

International safety regulations such as EN 1127-1 consider ultrasound to be an ignition source. Currently, applications of ultrasound in explosive atmospheres have to comply with a threshold value of 1 mW/mm2. However, it is unclear as to how this intensity has to be measured and, therefore, this threshold value is poorly defined. Moreover, it is based on theoretical estimations in analogy to other ignition sources and there are no publications or significant records on these estimations. Within a research project at PTB, it has now been investigated experimentally in relation to worst-case considerations including airborne ultrasound, focused MHz ultrasound in liquids and acoustic cavitation. On the basis of the results of the research it is now possible to revise the current regulations and to specify measures for safe operation of ultrasonic applications in explosive atmospheres. In this context, for ultrasound coupled directly to gaseous atmospheres a new threshold value of 170 dB (re. 20 mPa) can be suggested, and for ultrasonic applications in liquids, an augmentation can be made to the threshold to 400 mW/mm2.

© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

According to international standards, ultrasound is considered to be an ignition source (EN 1127-1, 2011) with a threshold value of 1 mW/mm2. This threshold value was developed on theoretical estimations in analogy to other ignition sources in a conservative way and, thus, includes a large safety margin. However, this threshold value is disputable and not well accepted by operators and manufacturers of ultrasonic equipment. This is because, on the one hand, it is not well defined from the metrological point of view. On the other hand, neither explosion accidents nor any experimental work related to ultrasound as an ignition source have ever been reported. Therefore, it was the goal of a research project at PTB to verify the capability of ultrasound to really ignite explosive atmospheres, i.e. to be incentive. Based on the theory of ultrasound and ignition processes, worst-case scenarios were developed and experimentally investigated. For the first time, it could be shown that airborne as well as liquid-borne ultrasound is really capable of setting off explosions. However, for ignition, an acoustically absorbing target object is needed to transform the acoustic energy

* Corresponding author. E-mail address: Michael.Beyer@ptb.de (M. Beyer).

into heat (Simon et al., 2014, 2015). The investigations on airborne ultrasound and on focused ultrasound at MHz frequencies are only briefly summarized in this paper and previously published papers are referred to. The scope of this paper encompasses the studies on the incendivity of acoustic cavitation and safety-related conclusions that can be drawn as a bottom line of the project. Therefore, requirements for applications of ultrasound will be reassessed and discussed.

Since at the starting point of the research project no elaborate investigations of ultrasound as an ignition source existed, theoretical considerations were taken into account to develop worst-case scenarios that could provoke ignition. This was done based on the literature on finite amplitude or high power ultrasound and on ignition mechanisms of different ignition sources (particularly hot spots and optical radiation) as well as on discussions with manufacturers of ultrasonic equipment, such as ultrasonic cleaning baths or level meters. In further steps, these worst-case scenarios were transformed into experimental setups. In pretests, they were verified and sharpened before ignition tests were conducted. In this way, the ignition tests addressed airborne ultrasound and liquidborne ultrasound with an explosive atmosphere above the liquid surface. In the case of liquid-borne ultrasound, a distinction was made between focused ultrasound at MHz frequencies and ultrasound at kHz frequencies that excited strong acoustic cavitation.

http://dx.doi.org/10.1016/jjlp.2015.03.010

0950-4230/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

2. Incendivity of airborne ultrasound

The ignition tests on the incendivity of airborne ultrasound are elaborately presented in (Simon et al., 2014). Hence, in this article, the results are only briefly summarized. For experimental studies on ultrasound as an ignition source, a suitable ultrasonic source for maximum sound pressure levels was needed. Especially for airborne ultrasound, technical and physical limits are set by the high impedance discontinuity between the solid sound source and the gaseous medium (Simon et al., 2013a, 2014; Gallego-Juárez et al., 2010). In addition, because of the wavelength of the order of 1 cm at frequencies typically in the kHz range, the focus area is limited to a width of this order as well. However, in resonant ultrasound standing wave fields high sound pressure levels can be attained efficiently. This can be achieved by a reflector positioned at a distance of a multiple of one-half wavelength. As sound sources, sonotrodes are suitable which act as sound particle velocity transformers and are employed, e.g. in ultrasonic levitators or ultrasound standing wave atomizers (Lessmann, 2004; Hemsel et al., 2005).

Ignition due to direct absorption of the sound wave is not possible (Simon et al., 2014) because of the limited sound pressure levels and the low acoustic absorption coefficients in gases, vapors or dusts (Bhatia, 1996; Dain and Lueptow, 2001; Ejakov, 2003; Bass et al., 1990; Lyman, 1977; Dukhin and Goetz, 2002). The situation changes, however, if (solid) materials with high acoustic absorption coefficients (above 5 kHz) are placed in the standing wave field and transform the acoustic energy into heat. In that case, the ultra-sonically heated target could lead to ignition at its hot surface.

According to these considerations, the ultrasound generating unit of an ultrasonic standing wave atomizer consisting of an electromechanical transducer, a booster and a sonotrode was used with an operating frequency of 20 kHz. The sonotrode's sound-emitting surface faced a reflector of steel. The distance between the sonotrode's surface and the reflector could be varied within the range of several wavelengths and, for ignition tests, was set to one wavelength. For temperature measurements, a thermocouple was integrated into the target's center while the sound pressure level was measured by a broad-band piezoelectric pressure sensor whose active measuring surface was integrated into the reflector's surface as illustrated in Fig. 1 (Simon et al., 2014). The sound pressure level is measured in dB as the sound pressure p relative to a reference value of p0 = 20 mPa: Lp = 20-log(p/p0).

The pretests showed that porous materials with open pores, such as foams and fibrous materials (mineral fibers, cotton), attained the highest temperatures within the ultrasound field. This is compatible with the theory on acoustic absorption in fibers (Mechel, 1989; Wang and Torng, 2001; Cuiyun et al., 2012). Furthermore, the position of highest heating rates was attained

Fig. 1. Setup for sound pressure measurement.

when a resonant standing wave field was generated and the target was positioned in its sound pressure anti-nodes (i.e. the positions of maximum sound pressure). This also is in accordance with theoretical estimations (Nyborg, 1981; Simon et al., 2014).

The ignition test yielded ignitions above a sound pressure level of 178 dB ± 2 dB. The target used (made of alkaline-earth silicate wool) attained temperatures between 250 °C and 500 °C, as dusts, sulfur (standard autoignition temperature (AIT) 250 °C), maize starch (AIT 380 ° C), calcium stearate (AIT 400 °C), aluminum (AIT 600 °C) and magnesium (AIT 590 ° C) were used. However, only sulfur dust could be ignited. Moreover, as vapors, carbon disulfide, diethyl ether, n-pentane and n-heptane were investigated, only in the cases of carbon disulfide was an explosion observed at a sound pressure level above 180 dB and a target temperature exceeding 400 °C.

3. Incendivity of liquid-borne ultrasound

In spite of similar principles of propagation, airborne and liquid borne ultrasound have to be considered separately in relation to ignition mechanisms. First, in the case of liquid-borne ultrasound, the ultrasound wave has to penetrate the liquid—gas phase interface in order to come into contact with an explosive atmosphere. At this interface, however, 99% of the sound wave will be reflected back into the liquid. Second, effects have to be considered that can be neglected for ultrasound in gaseous media or that only occur in liquids. On the one hand, because of lower absorption at MHz frequencies in liquids than in gases, it is possible to sharply focus the ultrasound to a beam width at focus of the order of 1 mm or less. Thus, high intensities can be easily attained. In contrast to sound waves in gaseous media, the intensity is defined in relevant standards, and, consequently, is uncritical for use as a threshold value. On the other hand, especially at kHz frequencies, acoustic cavita-tion is excited by the ultrasound in liquids, i.e. microbubbles oscillating in the ultrasound field grow and, at a certain point, vigorously collapse. At the moment of collapse extreme temperatures and pressures are attained inside the bubbles, so the question is raised as to whether they can cause ignition of an explosive atmosphere above the liquid surface.

3.1. Incendivity of focused ultrasound at liquid surfaces

The elaborate presentation of ignition test with focused ultrasound at liquid surfaces is given in (Simon et al., 2015) while in this article the results are only briefly summarized. Due to the possibility of sharply focusing MHz ultrasound in liquids, intensities of multiple W/mm2 can be attained easily when concavely shaped transducers are used and, thus, these intensities significantly exceed the current threshold value of 1 mW/mm2. An example is HIFU transducers (HIFU: high-intensity focused ultrasound) that are employed in therapeutics for the obliteration of tumors. In this case, the heating resulting from the acoustic absorption of the ultrasound by the tumor tissue is made use of.

Despite the reflection of the ultrasound wave at the liquid surface, ignition could yet occur if the acoustic energy is transferred into heat by a sound-absorbing solid fixed at the liquid surface adjacent to an explosive atmosphere. For this mechanism, however, the solid material has to show specific characteristics (Simon et al., 2015). First, its acoustic impedance has to be of the order of the liquid, so the sound wave is transmitted into the solid. This precondition already limits the range of suitable materials tremendously. Metals and ceramics which can endure temperatures exceeding the AIT of gases or vapors have acoustic impedances that are several orders higher than those of liquids so they can be excluded for this possible ignition mechanism. Data on the

acoustic properties can be found in the literature, for instance in (Cheeke, 2012, Ensminger, 1988) or (Deutsch et al., 1997). In addition to a matching acoustic impedance, the target material must show acoustic absorption in order to transform the acoustic energy into heat. Otherwise, the sound wave transmitted into the target will, again, be reflected at the phase interface between the target and the adjacent explosive atmosphere and run back into the liquid. The preconditions of the acoustic impedance and acoustic absorption are met by many plastics, for example PMMA. Furthermore it is known from experience that PMMA melts in the focus of HIFU transducers, which implies that it meets all preconditions for transferring acoustic energy into heat. However, its melting point is only around 150 °C and that is, therefore, too low to ignite vapor or gas and air mixtures. Yet there are plastics specifically designed for high temperature applications that endure temperatures of several hundred degrees such as polyetheretherketone (PEEK) with a melting point of approx. 350 °C. Therefore, this material was chosen as a target material for ignition tests. In addition, gypsum and graphite were considered as target materials because they also show acoustic impedances of the order of liquids. However, they yielded only little temperature augmentations in the focus of a HIFU transducer (Simon et al., 2015).

Ignition tests yielded ignitions of carbon disulfide (AIT 95 °C (Brandes and Möller, 2008)) and diethyl ether (AIT 175 °C (Brandes and Möller, 2008)) at the most easily ignitable mixture concentration according to (Welzel, 1996). The intensities had to exceed ¡sata = 528 mW/mm2 ± 8% (Isata: spatial-average temporal average intensity (IEC 62127-1, 2007)). At the threshold of Isata = 528 mW/ mm2, the time to ignition after starting insonification was in the range of approximately 1 min—2 min. At higher intensities, the time to ignition decreased to a few seconds. The intensity was determined by setting the acoustic power measured by an acoustic radiation force balance in relation to the area of the focused ultrasonic beam. The beam width at focus was determined by a membrane hydrophone and was defined as the decay of the intensity to 10% of the maximum signal (i.e. -10 dB). Below this intensity the heating rate was too low to attain hot surfaces above 150 °C. Thus, ignition at the hot surface can be excluded (Simon et al., 2015, 2013b).

In addition to the requirements concerning the material properties of the insonified target, it has to be fixed at the liquid surface in order to prevent it from sinking or from being pushed out of focus by the acoustic radiation pressure. Since the acoustic energy is absorbed inside of the target, its geometric dimensions have to be of the order of at least one wavelength. Otherwise, the sound wave will pass through the target without significant absorption and heating. Besides, the cooling by the liquid might dominate the ultrasonic heating (Simon et al., 2015).

3.2. Incendivity of acoustic cavitation

3.2.1. Theoretical considerations

At kHz frequencies, the wavelength of the ultrasound is of the order of several centimeters (7.5 cm at 20 kHz, 1.5 cm at 100 kHz). Therefore, sharp focusing is not possible. However, at these frequencies, strong acoustic cavitation can be observed. Acoustic cavitation is the formation and oscillation of bubbles in an insoni-fied liquid (Suslick and Flannigan, 2008). It is excited in the negative pressure phase of the ultrasound wave and can be subdivided into vaporous cavitation, when the cavities are formed under the influence of the negative sound pressure wave and, thus, contain only vapor from the surrounding liquid or gaseous cavitation when existing gas bubbles are excited to grow and collapse (Neppiras, 1980). Stable acoustic cavitation can be distinguished from transient cavitation. In the former case, the bubble motion is linear and

the bubbles grow in the rarefaction phase and decrease in the compression phase. The bubbles exist for several cycles of the ultrasound wave. Time scales for growth and compression are long enough to enable mass transfer and thermal diffusion can occur (Neppiras, 1980). In contrast, transient acoustic cavitation occurs at elevated sound pressure levels and the bubble motion becomes highly nonlinear. The bubbles oftentimes only exist for less than one cycle of the ultrasonic wave. After attaining diameters of the order of 100 mm they rapidly collapse to diameters of the order of 1 mm in about 1 ms in an adiabatic way and, thus, temperatures of several 1000 °C and extreme pressures are attained (Neppiras, 1980; Suslick and Flannigan, 2008). However, these conditions relate only to the last phase of the collapse with a duration of only several 1 ns and are often accompanied by light emission known as sonoluminescence (Suslick and Flannigan, 2008).

Considerations must be given to the fact that forces in the ultrasound field and between oscillating bubbles cause the bubbles to build clusters whose shape is determined by the sound pressure field (Mettin, 2005). Consequently, in the liquid there are clouds of bubbles potentially filled with explosive gases and, at the moment of bubble collapse, with multiple, extremely hot spots that are, however, very small and exist only for several nanoseconds. So the question is raised as to whether in these bubble clusters ignitions could occur that might advance into an explosive atmosphere above the liquid surface of an ultrasonic bath.

In order to investigate the incendivity of acoustic cavitation within a worst-case scenario, parameters have to be identified that might provoke an ignition. On the one hand, it is reasonable to excite strong acoustic cavitation and, on the other hand, to get it as close as possible to the liquid surface and, thus, as close as possible to an explosive atmosphere above this surface. Otherwise, a thick layer of liquid above a possible ignition in a cloud of cavitation bubbles might extinguish such an ignition before it can penetrate the explosive atmosphere. Furthermore, bubbles filled with an easily ignitable explosive gas mixture must be generated in the liquid.

In (Apfel, 1981) the requirements for the excitation of acoustic cavitation are described. The maximum temperature inside the collapsing bubbles depends on the gas inside the bubble and on the ratio of heat capacities g = cp/cV. Therefore, argon yields maximum temperatures (Flint and Suslick, 1989). Following (Suslick, 1988), acoustic cavitation is influenced by the frequency of the ultrasonic wave, the sound pressure amplitude, the temperature of the liquid, the amount of gas dissolved in the liquid as well as contaminants or particles suspended in the liquid. At low frequencies, the diameter of the cavitation bubbles increases while the amount of bubbles decreases. Increasing temperatures and sound pressure amplitudes yield larger bubbles. The number of bubbles increases with the amount of gas dissolved and of suspended particles in the liquid. According to 22, the most dense clusters of cavitation bubbles can be observed in front of high power sonotrodes. The investigations of (Nowak et al., 2009) imply that highest collapse speeds and, hence, temperatures are attained in large bubbles.

From the above follows that the worst-case scenario consists of a high-power sonotrode with an operating frequency of 20 kHz (lower barrier frequency of ultrasound) whose sound-emitting surface faces the liquid surface adjacent to an easily ignitable explosive atmosphere. The explosive gas mixture is injected into the ultrasonic field so, on the one hand, seeds for the generation of cavitation bubbles are supplied and, on the other hand, larger bubbles of several millimeters in diameter come into direct contact with the hot collapsing bubbles. At the same time, the continuous injection of explosive mixture into the liquid sustains the explosive atmosphere at the liquid surface. This worst-case situation is shown in Fig. 2 where a cannula is used to inject the explosive gas mixture.

3.2.2. Experimental investigations

The experimental investigations of the incendivity of acoustic cavitation were divided into two steps. In the first step, the worst-case situation described in Section 3.2.1 was transferred into an experimental setup and pretests were conducted to specify the conditions that might provoke an ignition. In the second step, ignition tests were conducted. The experimental setup is shown in Fig. 3. In the center of the setup a 20 kHz sonotrode (Telsonic AG, Sonoprozessor DG 2000) with a maximum power of 2000 W faces the liquid surface. The explosive atmosphere at the liquid surface was premixed in a gas mixture preparation in order to precisely set the mixture concentration. To check this concentration, a paramagnetic oxygen analyzer (Siemens Oxymat 6F) was used. From the concentration of oxygen in the vessel the concentration of combustibles could be determined. For the injection of explosive gas mixtures into the cavitation field in front of the sound-emitting surface, a cannula or, alternatively, a frit was used whose positions could be manually adjusted. At the top of the vessel a bursting foil served as a vent in case of possible ignition. As a liquid, water was used. This was done following the investigations of (Sychev and Pinaev, 1986), who investigated detonation waves in liquid-bubble systems and also used water as a liquid. The two phase system can be regarded as a medium with discretely distributed sources of energy. The crucial factor for ignition is the content of the bubbles rather than the liquid carrying these bubbles. Therefore, results are expected to be the similar if flammable liquids where used. The advantage of using water is that the sound pressure level could be easily measured by a hydrophone (Reson, type TC4013-1). If different liquids were used, the hydrophone had to be calibrated for each liquid. Also, it is to be avoided that especially aggressive solvents harm the hydrophone.

To monitor the inside of the vessel, a high speed camera and, alternatively, a reflex camera as well as a thermographic camera were used. The thermographic camera was focused on the bubbles ascending to the liquid surface to detect hot bubbles which would indicate bubble ignition below the liquid surface. Furthermore, the temperature of the liquid and of the explosive atmosphere above the surface were measured by two thermocouples.

Within the pretests, the level of liquid above the sound-emitting sonotrode was varied. The explosive mixture was injected by a frit with a pore size of 200 mm to "seed" a large amount of bubbles filled with explosive mixture into the sound field. The frit was then exchanged for a cannula to inject explosive bubbles directly into

Fig. 2. Worst-case situation for the investigation of the incendivity of acoustic cavitation.

Fig. 3. Worst-case situation for the investigation of the incendivity of acoustic cavitation.

spots of strong cavitation. The position of the frit and the cannula was manually adjusted to achieve good interaction between the injected bubbles and the cavitation bubbles. With respect to the ultrasound signal, the power was varied and the operation mode changed between continuous wave and pulsed mode with different pulse lengths.

In accordance with the results of the pretests ignition tests were conducted. The explosive atmosphere and the mixture injected into the liquid were premixed by the gas mixture preparation. As explosive atmospheres, carbon disulfide (2%—7%) and air, diethyl ether (12%) and air, and hydrogen (24%) and air were used. The vessel was flushed with the explosive mixture until the desired concentration was reached. By a junction in the supply it was then switched into the injection device to aerate the water with the explosive mixture. To sharpen the worst-case conditions further, ignition tests with hydrogen-air as an explosive atmosphere were conducted while injecting the hydrogen—oxygen mixture (in the ratio 2:1) into the liquid. For injection, the cannula as well as the frit were used.

The sonotrode and the excited cavitation heated the water. Therefore, a cryostat was used during ignition tests with diethyl ether to keep the water temperature constant. However, the tests with carbon disulfide and, in part, the tests with hydrogen—oxygen, were carried out without the cryostat to find out whether the temperature rise promoted ignition. In general, though, it is assumed that ignition would be triggered within the first minutes of insonification because of the large number of cavitation events per second. For this reason, most of the ignition tests with hydrogen-air and hydrogen—oxygen were conducted only for a few minutes.

3.2.3. Results

The results of the hydrophone measurements in front of the sonotrode are shown in Fig. 4. This measurement gives an idea of the magnitude of the sound pressure level; however, the incident sound wave from the sonotrode interferes with the reflected wave from the liquid surface. Besides, each imploding cavitation bubble also sends out its own sound wave. To reduce the influence of the reflection from the liquid surface, the measurement was done at a water level of 75 mm over the sound-emitting surface, whereas

liquid surface

Fig. 4. Results of the hydrophone measurements in front of the sonotrode.

during ignition tests water levels of 40 mm at maximum were used. The measurement shows that the maximum sound pressures are attained in the region from the top of the sonotrode to 35 mm. At larger distances, the sound pressure rapidly decreases.

In Fig. 5 the influence of the water level on the formation of the acoustic cavitation field is shown. At low water levels, the acoustic radiation pressure yields a bulge forming on the liquid surface and droplets are sprayed into the gas phase. The higher the water level, the calmer the liquid surface and the more distinct the field of cavitation bubbles is. Picture 4a) of the sequence in Fig. 5 depicts how the cavitation bubbles move in the shape of an arch and merge in its peak.

The impact of the cavitation on the injected bubbles is illustrated in the sequence of Fig. 6. While the ultrasound was turned off, the injected bubbles showed a diameter of about 1 mm. After turning on the ultrasound, a cloud of small cavitation bubbles rose and atomized the injected bubbles into smaller ones. This reduces the chances for ignition as the investigations of (Mitropetros, 2005) on bubble-liquid systems show.

The sequence in Fig. 7 shows the interaction of acoustic cavi-tation with the bubbles injected by the cannula. The injected bubbles had diameters of about 2 mm—3 mm. Moreover, it can be seen how the cavitation bubbles merge directly beneath the liquid

surface and how the acoustic radiation pressure displaces the stream of injected bubbles to the liquid surface. Therefore, for most ignition tests, the explosive mixture was injected into or just below the merging point of the cavitation bubbles because it is assumed that this is the point of strongest cavitation activity. In addition, only a thin layer (about 5 mm) of water separated this point from the explosive atmosphere at the liquid surface.

The employed power sonotrode can operate at a maximum electric input power of 2000 W. However, because of the impedance discontinuity between the sonotrode and the liquid-bubble mixture only input powers of 650 W could be used. Higher power levels can only be coupled into the water if the vessel is set under a static pressure such that this parameter is maximized. Ignition tests were conducted in the continuous wave operation mode of the sonotrode as well as in the pulsed mode. It showed that pulse lengths of 1 s—2 s were appropriate, since at shorter pulse length only weak and unstable cavitation fields were formed.

None of the ignition tests resulted in ignition of the explosive atmosphere above the liquid surface. Furthermore, there was no evidence that ignition occurred beneath this surface that was extinguished by the water layer before it could penetrate the atmosphere. However, during ignition tests with carbon disulfide the water grew yellow and cloudy and small yellow (in the web version) particles could be observed. This shows, beyond the sequence in Fig. 6, that the cavitation bubbles interact with the injected bubbles.

4. Discussion of results of acoustic cavitation

Despite the extreme conditions inside the collapsing cavitation bubbles, they are not incendive with respect to explosive atmospheres. The experiments modeled worst-case situations bringing strong cavitation clusters into direct contact with explosive atmospheres of the most critical gases and vapors at their most easily ignitable concentrations. The cavitation activity was maximized by first using the maximum power that could be coupled into the liquid, by using a frequency of 20 kHz which produces the most violent cavitation (cf. Section 3.2.1) and, second, by continuously feeding gas-air and vapor-air bubbles respectively, into the liquid. The high speed recordings ensure that the cavitation bubbles really interact with the injected bubbles filled with explosive mixture. However, this interaction led to an atomization of these bubbles rather than to ignition that could propagate into the explosive atmosphere above the liquid surface. The interaction between cavi-tation and injected bubbles is, again, underlined by the fact that during the test with carbon disulfide sulfur particles were produced. This shows that chemical reactions were excited inside the

field of cavitation

Fig. 5. Influence of the height of the water level over the sonotrode. Water levels 1) and 1a): 0 mm, 2) and 2a): 7 mm, 3) and 3a): 23 mm, 4) and 4a): 33 mm.

Fig. 6. Impact of the acoustic cavitation on the injected bubble: 1) Ultrasound turned off 2) t = 0: Ultrasound turned on 3) t = 0.035 s: Excitation of cavitation bubbles 4) t = 0.08 s and 5) t = 0.134 s: Destruction of injected bubbles under the impact of cavitation bubbles.

bubbles on account of this interaction. However, this reaction did not result in ignition. Finally, the fact that not explosion occurred even in the ignition test with hydrogen—oxygen shows that acoustic cavitation cannot ignite a hydrogen—air mixture even with a larger number of ignition tests. Also, it ensures that this mechanism is not incendive even if the setup used has not modeled the worst-case situation 100%.

The results are supported by the investigations of (Nguyen and Jacqmin, 2005) concerning the so-called "cavitation bubble combustion" which has been observed in hydraulic systems (Lohrentz, 1968). In the research conducted by (Nguyen and Jacqmin, 2005) two phase systems of methanol-air and water-methane were investigated under the influence of strong cavitation. In a similar way, it was concluded that chemical reactions are excited by the acoustic cavitation, however, explosions could not be observed.

The ineffectiveness of acoustic cavitation to cause ignition of explosive atmospheres can be attributed to various effects. First, the extreme temperatures and pressures inside the cavitation bubbles are attained only in the last phase of their collapse (Suslick and Flannigan, 2008). In this phase, their diameter is of the order of 1 mm or even less and this phase lasts only for a few nanoseconds (Suslick and Flannigan, 2008). Because of the compressed "gas cushion", the bubble immediately rebounds again. Warnatz, (1993) determined ignition delay times of the order of 10 ms for hydrocarbon-air mixtures at temperatures of2000 K. This is at least two orders of magnitude longer than the critical phase during the bubble collapse. Together with the surrounding liquid, the rebound of the bubble leads to cooling rates of up to 1012 K/s (Suslick and Flannigan, 2008). This implies that the cavitation bubbles, on the one hand, cannot ignite themselves as the phase of the collapse is too short and, on the other hand, they are too small to ignite the injected bubbles.

5. Safety-related requirements

In the previous section's theoretical considerations, experimental investigations and results of ignition tests concerning the incendivity of ultrasound were presented. Based on these findings, the ignition source ultrasound can be much more precisely described and the threshold value for the application of ultrasound in explosive atmospheres can be significantly augmented.

In contrast to the current threshold value of 1 mW/mm2 that is valid for all applications of ultrasound (EN 1127-1, 2011), it follows from the theoretical considerations in Section 2 and Section 3.1 that a distinction should be made between airborne and liquidborne ultrasound. While in the latter case the acoustic intensity is well defined in specific standards (cf. Section 3.1), in the former case it is more appropriate to measure the sound pressure level (cf. Section 2). For this reason, in the following, requirements for the safe application of ultrasound in explosive atmospheres are presented which are subdivided into general requirements as well as requirements specifically developed for airborne and liquidborne ultrasound. For each of the three cases the requirements are first presented and subsequently explained. The enumeration of requirements, however, is maintained throughout the three parts.

5.1. General requirements

5.1.1. Presentation of the requirements

1. Up to a frequency of 10 MHz, ultrasound can only indirectly ignite explosive atmospheres via absorption of the ultrasound wave by a sound-absorbing solid. Direct ignition via absorption of the acoustic energy by the explosive atmosphere is not possible at these frequencies.

Ultrasound of frequencies above 10 MHz was not considered in the research conducted. Hence, the limitations according to currently valid regulations (e. g. EN 1127-1, 2011) have to be maintained.

2. Sound-absorbing materials which can easily catch fire (such as cotton i.e. at kHz frequencies) should always be avoided in strong ultrasound fields, independent of the kind of explosive atmosphere, because the burning absorber could ignite an explosive atmosphere.

5.1.2. Discussion of the requirements

In item 1, the area of validity for the following requirements is defined. As lower limiting frequency 20 kHz is regarded in

Fig. 7. Injection of explosive mixture by the cannula into the merging point of the cavitation bubbles close to the liquid surface. 1) Ultrasound is turned off, bubbles are injected. 2) Ultrasound is turned on, acoustic cavitation is excited. The injected bubbles are displaced by the acoustic radiation force. 3) and 4) Cavitation bubbles show a stable figure with a merging point beneath the liquid surface. The bubbles with explosive mixture are injected into this merging point.

accordance with the literature (Cheeke, 2012). Above 10 MHz, effects such as molecular resonance might become prominent, which have not been considered in the present studies. Consequently, it is only possible to make general statements concerning frequencies in the frequency range from 20 kHz to 10 MHz. In this range, ignition is only possible by generation of hot surfaces on solids in the ultrasound field, which show high acoustic absorption at kHz frequencies and above.

Item 2 refers to the case of a sound-absorbing solid catching fire as a result of the ultrasound-induced heating. Consequently, the open flame could ignite the explosive atmosphere. Critical materials are cotton wool or wadding, in particular.

5.2. Requirements for airborne ultrasound

5.2.1. Presentation of the requirements

3. Ultrasound cannot effectively ignite explosive atmospheres of dusts if the sound pressure level in the whole sound field is below 170 dB (re. 20 mPa). This is due to the fact that even in solids with 100% sound absorptivities (at 5 kHz), no critical heating is attained.

This threshold comprises a safety margin of 6 dB and takes into account a measurement uncertainty of 2 dB (k = 1).

4. In applications that do not comply with this threshold, fixed sound-absorbing bodies must be barred from insonification periods longer than 1 s.

5. The threshold of 170 dB (re. 20 mPa) is also applicable to explosive atmospheres of gases and vapors because critical temperatures for these atmospheres are not reached below this threshold.

5.2.2. Discussion of the requirements

Beyond the general requirements for all ultrasound fields, item 3 through item 5 refer to applications where the ultrasound is coupled directly into the explosive atmosphere. In contrast to the current threshold value of 1 mW/mm2, item 3 proposes a sound pressure level instead of an intensity for airborne ultrasound. In gaseous mediums the intensity is not well defined. The sound pressure level can be easily measured by a calibrated dynamic broadband pressure sensor, whereas the sound intensity I has to be calculated as the product of the effective values of sound pressure peff and particle velocity veff, I = peff-veff. Moreover, in general, sound pressure and particle velocity are out of phase and, thus, the phase angle has to be determined. In addition, in a standing wave field this phase angle is zero, which means that the intensity disappears even though the research showed that this really is the most critical case (Simon et al., 2014).

If the sound pressure level does not exceed 170 dB throughout the whole sound field, ignition is not possible. In this case, as the experiments conducted showed, even materials specifically designed to absorb sound waves at 5 kHz or even higher do not heat up to temperatures above 200 °C, which would be critical for ignition. However, it has to be ensured that the sound pressure level is below this value in every spot of the sound field. Critical spots are particularly sound pressure anti-nodes of standing wave fields, the focus of a concavely shaped stepped sonotrode (Gallego-Juárez et al., 2010), or the near field of a transducer (Cheeke, 2012). These three critical situations are illustrated in Fig. 8a) through c). Accordingly, the sound pressure should be measured at the surface of a reflector at a distance of one wavelength from the sound-emitting source (Fig. 8a)), at the near field border (Fig. 8b)), or in

case of a concavely shaped transducer, in the focal area (Fig. 8c)). The near field border is the distance N = D2 - A2/4AzD2/4A, where D is the active diameter of the transducer and l the wavelength in the gaseous medium.

Item 4 relates to the situation where the sound pressure exceeds the threshold value at some point of the sound field. Then, it has to be ensured that no highly sound-absorbing bodies that are fixed in the sound field are exposed to these sound pressure levels which might yield hot spots capable of igniting the explosive atmosphere. Especially critical are porous materials with open pores (e.g. foams) or fibrous materials (e.g. mineral wools), because they show high absorptivity at a frequency of 5 kHz and above.

Item 5 addresses the threshold in relation to explosive atmospheres of gases and vapors. During ignition tests with such atmospheres only carbon disulfide yielded ignition. Despite its lower AIT (95 °C (Brandes and Möller, 2008)) compared to the sulfur dust used (AIT 250 °C), ignition only occurred at sound pressure levels (>180 dB re. 20 mPa) and also at temperatures (>400 ° C) higher than for sulfur dust. Therefore, the recommended threshold of 170 dB (re. 20 mPa) is valid for dusts as well as for vapors and gases.

5.3. Requirements for liquid-borne ultrasound

5.3.1. Presentation of the requirements

6. Ultrasound cannot ignite explosive atmospheres above a liquid surface if the acoustic intensity at the liquid surface does not exceed 400 mW/mm2, because no critical temperatures can be induced in sound-absorbing solids that penetrate the liquid surface. This threshold comprises a safety margin of 20% in relation to the experimentally determined ignition limit for diethyl ether and takes into account a measurement uncertainty of 8% (k = 1). The threshold value is representative of all gases and vapors, including carbon disulfide.

7. Compliance with the threshold can be confirmed for application in practice in the following way.

a. For single ultrasound sources

- with an active sound-emitting diameter greater than or equal to the sound wavelength in the liquid, the output power of the source set in relation to the square of the wavelength in the liquid must not exceed the threshold value (Fig. 9, Case 1). It is sufficient if this condition is in compliance with the electric input power of the source.

- with a diameter less than the ultrasound wavelength in the liquid the acoustic output power set in relation to the sound-emitting surface of this source must not exceed the threshold (Fig. 9, Case 2).

b. For applications consisting of multiple transducers.

- where additive overlaying of the sound field maxima of the individual sources in the far field is possible, the sum of the intensities estimated according to a. may also not exceed the threshold (Fig. 10, Case 3).

- where the sound field maxima of the individual sources in the far field cannot additively overlay (Fig. 11, Case 4), it is sufficient to evaluate each of the sources separately according to a.

c. If the intensities estimated according to a. and b. exceed the threshold value, compliance with the threshold can, alternatively, be verified by determination of the sound pressure maxima via hydrophone measurements at the level of the liquid surface in the application: The intensity I calculated from the sound pressure p, I = p(t)2/(p-c), where p is the density and c the sound velocity of the liquid, must not exceed the threshold (cf. (IEC 62127-1, 2007)). The

Fig. 8. Critical spots to determine the sound pressure level in the case where a) a standing wave field, b) an unfocused propagating wave, and c) focused ultrasound could be generated.

Fig. 9. Distinction of cases according to the geometric dimensions of the sound-emitting source in relation to the ultrasound wavelength.

Fig. 10. Case 3: Additive overlay of the sound field maxima in the far field.

hydrophone used for the sound pressure measurement has to be calibrated and traced to reference standards.

In the case of spatially severely limited sound beams which can occur in the frequency range above 500 kHz, the threshold may be calculated as the intensity Isata (spatial-average temporal-average intensity, cf. (IEC 62127-1, 2007)) averaged over the beam's cross section (-12 dB beam width). Instead of the determination of Isata averaged over the sound pressure profile, the acoustic output power can be measured (cf. (IEC 61161-3, 2013)) which must be

Fig. 11. Case 4: No interference of the sound field maxima in the far field.

divided by the minimum -12 dB-area at focus.

8. In the case where the threshold value is exceeded, ignition has

to be excluded by different means. Such means can be:

- exclusion of acoustically absorbing bodies that penetrate the liquid surface and have an acoustic impedance similar to liquid,

- exclusion of acoustically absorbing bodies that are spatially fixed and penetrate the liquid surface,

- limitation of the insonification time of acoustically absorbing bodies that penetrate the liquid surface.

5.3.2. Discussion of the requirements

The requirements of item 6 through item 8 address applications where ultrasound is coupled to a liquid that is in contact with an explosive atmosphere. In item 6, a threshold value of 400 mW/mm2 is proposed for such applications. This means an augmentation of the current threshold by a factor of 400, while a safety margin of 20% is taken into account. This is because no research has been done to optimize the target material. However, this safety margin is regarded as sufficient, since multiple conditions have to be met simultaneously for ignition:

• First, the intensity of the incident sound wave must exceed the threshold mentioned above.

• Second, an acoustically absorbing solid must be fixed at the liquid surface which has an acoustic impedance that matches the impedance of the liquid and has a high absorption coefficient. Also, its geometric dimensions in the direction of propagation of the sound wave must be large enough for absorption (i.e. several millimeters at least). In addition, it must be temperature resistant in order to attain temperatures exceeding the AIT of gases and vapors. As target materials, contaminants floating to the surface or components in a cleaning bath have to be considered that match the characteristics mentioned above. Critical materials are, in particular, viscoelastic materials such as plastics or bitumen (Simon et al., 2015).

• With respect to the explosive atmosphere at the liquid surface the most easily ignitable mixtures of diethyl ether and carbon disulfide were used during ignition tests. At the position of a target in real-life applications, the mixture concentration will significantly differ from the most easily ignitable concentration in most of the cases.

Item 7 describes possible ways to verify that a given application complies with the threshold value. In item 7 a. two marginal cases are distinguished. This is, on the one hand, a focusing transducer and, on the other hand, a point-like sound source (Fig. 9). In the first case, it is necessary for sharp focusing that the diameter D of the transducer be larger than or at least equal to the wavelength l of the ultrasound. Furthermore, it is assumed that the ability to focus the sound beam is limited to the order of the wavelength. Therefore, the focal area can be estimated to be AFocus = l2 and, in turn, the intensity can be appraised by P/l2, where P is the acoustic output power of the transducer. This, however, is an overestimation on the safe side, i.e. the intensity is in reality lower.

The second case in Fig. 9 refers to a point-like sound source which has a diameter that is less than the wavelength: D < l. Accordingly, the sound field diverges and the maximum intensity is attained directly in front of the sound-emitting surface. Thus, it is sufficient to set the acoustic output power in relation to the active sound emitting surface P/Aeff. In both cases, the electrical input power can be used to estimate the intensity if the acoustic output power is unknown.

If the sound field is generated by more than just one source, item 7 b. is applicable (cf. case 3 in Fig. 11 and case 4 in Fig. 11). In this case, it is necessary to check whether the interfering sound fields can superpose additively and consequently exceed the threshold value, even though the intensities of each source are below this limit. If the arrangement of the sources leads to an interference of the sound field maxima in the far field (cf. Fig. 11), in the first step, the intensity of each sound source can be estimated separately according to item 7 a. In the second step, these intensities have to be added and the sum must not exceed the threshold value. In contrast, if such interference can be excluded, it is sufficient to only evaluate each ultrasound transducer separately according to item 7

If all of the appraised intensities are below the threshold, the compound system is regarded as being safe as well. If the estimated intensity according to item 7 a. and item 7 b. exceeds the threshold, item 7 c. states that it is possible to conduct hydrophone measurements in order to show that the intensity to be considered is in compliance with the limit. During these hydrophone measurements, however, it has to be considered that the measurement directly beneath the liquid surface is influenced by the reflection from the liquid-gaseous phase interface even though only the incident sound wave directly from the source is relevant. Accordingly, it is possible to determine the sound pressure at the level of the liquid surface in the application and to suppress the influence of the reflection during the hydrophone measurements, e.g. by a higher liquid level. Alternatively to measuring the intensity averaged over the sound pressure profile, a combination of an output measurement by an acoustic radiation force balance in combination with a hydrophone measurement of the beam width is possible. The measurements of the intensities in the present research project were performed in this way (Simon et al., 2015). The advantage is that the width of the beam can be measured at low output powers, while high powers could harm the hydrophone because of cavita-tion. In contrast, the acoustic radiation force balance is less sensitive to acoustic cavitation.

Item 8 allows the threshold value to be exceeded in cases where different measures are taken to prevent ignition by ultrasound. Since ignition is only possible by absorption of the ultrasound wave and heating of an ultrasound absorbing target fixed at the liquid surface it is, consequently, permitted to exceed the threshold value if such targets themselves or insonification times longer than 1 s can be excluded.

б. Conclusions

The present contribution gives an overview on the studies concerning the incendivity of ultrasound and the safety-relevant conclusions that can be drawn. The research shows that, on the one hand, ultrasound is incendive toward explosive atmospheres of dusts vapors and gases irrespective whether the ultrasound is coupled into a liquid or a gaseous medium. However, in both cases, the ultrasound has to be absorbed by a target material of specific characteristics which result in the heating of the target that could lead to an ignition at its hot surface. The theoretical considerations and experimental results that these findings are based on, were summarized. In addition, acoustic cavitation is regarded and the most significant theoretical considerations and experimental investigations comprising ignition tests based on worst-case conditions are presented. It follows that acoustic cavitation cannot trigger ignition because the collapsing cavitation bubbles are too small and the critical phase of the collapse is too short.

The results from ignition tests with airborne ultrasound in dusts and vapors and with liquid-borne ultrasound in gases and vapors show, however, that the currently valid threshold value fixed in

international standards is inappropriate. This is because, on the one hand, the intensity is not explicitly defined for ultrasound in gaseous media and, thus, it is more suitable to limit the sound pressure level which can be directly measured. On the other hand, the current threshold is too low and can be significantly augmented while keeping the same level of safety.

Thus, based on the research conducted in relation to worst-case considerations and ignition tests, new safety-relevant requirements for applications of ultrasound in explosive atmospheres could be developed which are meant as a recommendation for a reassessment of this ignition source for the relevant standards and regulations. These requirements comprise a threshold value of 170 dB (re. 20 mPa) for applications of airborne ultrasound and 400 mW/mm2 for ultrasound in liquids. On the bottom line, all currently known applications of ultrasound are safe or can be designed to be safe by simple measures. Furthermore, the recommended reassessment of requirements could open up innovative applications of ultrasound.

Acknowledgments

The authors gratefully acknowledge the support of the Deutsche Gesetzliche Unfallversicherung (German Social Accident Insurance).

References

Apfel, R., 1981. Ultrasonics, Chapter Acoustic Cavitation. Academic Press, pp. 355—407.

Bass, H., Sutherland, L., Zuckerwar, A., 1990. Atmospheric absorption of sound:

update. J. Acoust. Soc. Am. 88 (4), 2019—2021. Bhatia, A., 1996. Ultrasonic Absorption. Clarendon Press, Oxford. Brandes, E., Möller, W., 2008. Safety Characteristic Data. In: Flammable Liquids and

Gases, vol. 1. Wirtschaftsverlag NW. Cheeke, J., 2012. Fundamentals and Applications of Ultrasonic Waves, second ed. CRC Press.

Cuiyun, D., Guang, C., Xinbang, X., Peisheng, L., 2012. Sound absorption characteristics of a high temperature sintering porous ceramic material. Appl. Acoust. 73, 865—87 .

Dain, Y., Lueptow, R., 2001. Acoustic attenuation in a three-gas mixture: results.

J. Acoust. Soc. Am. 110 (6), 2974—2979. Deutsch, V., Platte, M., Vogt, M., 1997. Ultraschallprüfung: Grundlagen und industrielle Anwendungen. Springer Verlag, Berlin. Dukhin, A., Goetz, P., 2002. Ultrasound for Characterizing Colloids — Particle Sizing,

Zeta Potential. In: Rheology, vol. 1. Elsevier Science & Technology. Ejakov, S.G., 2003. Acoustic attenuation in gas mixtures with nitrogen: experimental data and calculations. Acoust. Soc. Am. 113 (4), 1871—1879. EN 1127-1, 2011. Explosive Atmospheres — Explosion Prevention and Protection — Part 1: Basic Concepts and Methodology.

Ensminger, D., 1988. Ultrasonics — Fundamentals, Technology, Applications. Dekker, Inc.

Flint, E., Suslick, K., 1989. Sonoluminescence from nonaqueous liquids: emission from small molecules. J. Am. Chem. Soc. 111, 6987—6992.

Gallego-Juarez, J.A., Rodriguez, G., Acosta, V., Riera, E., 2010. Power ultrasonic transducers with extensive radiators for industrial processing. Ultrason. Sono-chem. 17, 953—964.

Hemsel, T., Hennig, C., Kauczor, C., Littmann, W., 2005. Powder coating fabrication by ultrasonic standing wave atomizer. HNI Verlagsschriftenr. 180, 183—186.

IEC 61161-3, 2013. Ultrasonics — Power Measurement — Radiation Force Balances and Performance Requirements.

IEC 62127-1, 2007. Ultrasonics — Hydrophones — Part 1: Measurement and Characterization of Medical Ultrasonic Fields up to 40 MHz.

Lessmann, N., 2004. Numerical and Experimental Investigation of the Disintegration of Polymer Melts in an Ultrasonic Standing Wave Atomizer. Fakultöt für Naturwissenschaften, Universität Paderborn (PhD thesis).

Lohrentz, H., 1968. Die Entwicklung extrem hoher Temperaturen in Hydrauliksystemen und die Einflüsse dieser Temperaturen auf die Bauteile und ihre Funktionen. In: Mineralöltechnik, vol. 11. Mineralöl Zentralverband e.V.

Lyman, F., 1977. Attenuation of high-intensity sound in a droplet laden gas. J. Sound Vib. 51, 219—235.

Mechel, F., 1989. Schallabsorber Band I: Außere Schallfelder - Wechselwirkungen. Hirtzel Verlag.

Mettin, R., 2005. Bubble and Particle Dynamics in Acoustic Fields: Modern Trends and Applications, Chapter Bubble Structures in Acoustic Cavitation. Research Signpost, Kerala, India, pp. 1—36.

Mitropetros, K., 2005. Shock Induced Bubble Explosions in Liquid Cyclohexane. Technische Universitöat Berlin (PhD thesis).

Neppiras, E.A., 1980. Acoustic cavitation. Phys. Rep. 61, 159—251.

Nguyen, Q.V., Jacqmin, D.A., 2005. A Study of Cavitation-ignition Bubble Combustion. NASA Glenn Research Center. Technical report.

Nowak, T., Mettin, R., Lauterborn, W., 2009. Bubble Motion and Jetting at Sonotr-odes. Deutsche Gesellschaft für Akustik e.V. (DEGA), pp. 712—714.

Nyborg, W., 1981. Heat generation by ultrasound in a relaxing medium. J. Acoust. Soc. Am. 70 (2), 310—312.

Simon, L.H., Fedtke, T., Wilkens, V., Beyer, M., 2013a. Zündwirksamkeit von Ultraschall beim Einsatz in explosionsfaöhigen Atmosphaören. Mittl. 123 (3), 31—42.

Simon, L.H., Fedtke, T., Wilkens, V., Beyer, M., 2014. Ignition of dust-air atmospheres by ultrasonic waves. J. Loss Prev. Process Ind. 26, 1583—1589.

Simon, L.H., Meyer, L., WIlkens, V., Beyer, M., 2015. Ultrasonically triggered ignition at liquid surfaces. Ultrason. Sonochem. 22, 235—242.

Simon, L.H., Wilkens, V., Beyer, M., 2013b. Incendivity of ultrasound applied in explosive atmospheres. In: 24th International Colloquium on Dynamics and Explosion in Reactive Systems, Taipei.

Suslick, K., 1988. Ultrasound — its Chemical, Physical, and Biological Effects. VCH Publishers, Inc.

Suslick, K., Flannigan, D., 2008. Inside a collapsing bubble: sonoluminescence and the conditions during cavitation. Annu. Rev. Phys. Chem. 59, 659—683.

Sychev, A.I., Pinaev, A., 1986. Self-sustaining detonation in liquids with bubbles of explosive gas. J. Appl. Mech. Tech. Phys. 27 (1), 119—123.

Wang, C.N., Torng, J.H., 2001. Experimental study of the absorption characteristics of some porous fibrous materials. Appl. Acoust. 62, 447—459.

Warnatz, J., 1993. Resolution of gas phase and surface chemistry into elementary reactions. Proc. Combust. Inst. 24, 533—579.

Welzel, M., 1996. Entzündung von explosionsföahigen Dampf/Luft- und Gas/LuftGemischen durch kontinuierliche optische Strahlung. Technische Universitaöt Braunschweig (PhD thesis).