JNIOSH

Abstract of Special Research Report (RR-94)

National Institute of Occupational Safety and Health, Japan

A Construction Method of Safety Work System Based on Confirming Rotation Stop

RR-94-1
Noboru SUGIMOTO

: From a viewpoint of safety, such an operation for machine-care as trouble-shooting and maintenance may be referred as a man-machine work system in which machines and human workers work together in a common work area. In such a man-machine work system, an interlocking is usually constructed so that workers are permitted to work freely in the work area and safety must be assured only on the hazardous machine-side in such a way that while machines are working sudden intervention of the work area by human workers must force any machine operation to stop. However, in case that they have long run-down time after triggering stop control, it may be impossible to assure safety only on the machine-side.
   This paper presents a safety interlocking system as an example for assuring safety in working together in hazardous area. The interlocking system is based on confirming the stop condition of a machine after run-down time and permitting human to enter into the work area. A rotation-stop sensor for detecting the stop condition of a driving mortor is proposed. This sensor issues a work permission signal to a door switch to unlock the door-gate after confirming machine stop. And the method of constructing this sensor on fail-safe is presented. The signal indicating the rotation-stopstate confirmation is produced in a fail-safe manner, i.e. by a logical product of the confirmation signal reporting that the sensor monitoring rotation of the motor is in a normal condition and the output of NOT operation which never erroneously judges to be in a rotation-stop-state and never permits human workers to enter to the dangerous space.
   The safety system proposed in this paper meets the requirement of Section 9.7 of British Standard Cord of practice for safety of machinery.

On the Buckling Coefficients Used for the Calculation of Crane Structure

RR-94-2
Yutaka MAEDA

: In this report, the theory and the calculation procedure of the buckling coefficients used for the buckling strength calculation of crane structure were clarified and summarized at first. Buckling stress in the elastic range was calculated by Euler's formula, and that in the plastic range was by Jäger's formula. The buckling stress in the plastic range was obtained under the following assumptions; (a) steel had the complete elasto-plastic characteristics, (b) the compressive force acted to the column with the initial eccentricity a, where a /i = 1/20+λ/500, i was the radius of gyration of sectional area, and λ was the effective buckling length of the column, (c) the center line of the column deformed along a sine curve.
    Buckling stresses in both ranges were divided by the safety factors, 2.5 or 1.5 respectively. The allowable buckling stress was defined as the minimum of these values or the allowable compressive stress. The buckling coefficient ω was obtained as the quotient of allowable compressive stress divided by the allowable buckling stress. Allowable buckling stress of a general column was derived from a column of T-type section, and that of a pipe column was from the pipe section, whose shell thickness was 1/6 of the diameter. As both of the derived equations were irrational, the Newton-Raphson method was used to solve them numerically.
    Secondary, values of allowable compressive stress and yield point, which had been used to calculate the buckling coefficients in the present construction code for crane and mobile crane, were obtained by the reverse calculation, and it became clear that they had some differences from the provided allowable compressive stress of the steel concerned.
    Concerning with the changes of allowable stress of steel to be the smaller value of the yield point divided by 1.5 and tensile strength divided by 1.8, and the range of steel grade, it was discussed the safest condition of the values of yield point and tensile strength for the decision of the buckling coefficients in each new range of steel. It became clear that in the area of σB being greater than 1.2σe, the safest ƒÖ was obtained by the greatest σe, and in the area of σB being less than 1.2σe, it was by the least σe under the condition of the maximum σB. The new buckling curves of those steels were calculated.

A Basic Study on the Fatigue Crack Monitoring Using Strain Gages

RR-94-3
Tetsuya SASAKI

: Fatigue is a major factor which must be taken into consideration to prevent fracture accidents of industrial installations and structures. A lot of researches were devoted to find out a method to prevent crack initiation due to fatigue. These studies, however, revealed that it was vain to make efforts to completely prevent fatigue crack initiation. Then new design concept called damage tolerance design was proposed and it has been applied to the design of fatigue critical structures. The damage tolerance design is a synthetic method to ensure the integrity of structures during their service life. Nondestructive inspection (NDI) is applied to materials and structures to verify the required integrity by detecting and characterizing cracks and flaws. The damage tolerance design has contributed to the decrease of fracture accidents due to fatigue.
   Recent fracture accidents of airplanes, however, revealed that the damage tolerance design was also imperfect. No one can detect all cracks and flaws by NDI. Some cracks and flaws are not detected because their size is under NDI level which refers to the degree of resolution, and some are not detected due to human error in inspection. These problems of the damage tolerance design are rejected if all fatigue cracks which will result in a fatal accident can be always monitored. This idea may seem to be unrealistic. But, fortunately, it is not necessary to monitor all parts of the installation or structure. The place of crack initiation due to fatigue is almost limited to the stress concentrated part and it can be easily prospected.
   Thus, this paper treats a basic study to monitor fatigue cracks using strain gages. Strain gages are selected for the sensor in this study because they are small, light, durable and low cost enough to be applied to the real structure. Two methods are proposed to monitor crack initiation and propagation. One is a direct method which detects the change of strain near crack tip, and the other is an indirect compliance method which detects the change of compliance due to a crack. An automated data logging system using real-time multi-task processing is also developed to store all the data during fatigue crack growth tests. Following conclusions are reached by fatigue crack growth tests using CCT specimens:
  (1) Fatigue cracks can be monitored by the direct method if a strain gage is close enough to crack tip. But the resolution drastically drops if crack initiation does not occur near the strain gage.
  (2) Fatigue cracks can be monitored more successfully by the indirect compliance method. The method has sufficient resolution even if the place of crack initiation cannot be precisely prospected.

Evaluation of Loudness for Intermittent Noise Based on Personal Auditory Characteristics

RR-94-4
Yoshiyuki EGAWA

: The object of this report is to evaluate the loudness of many kind of intermittent noises based on personal auditory characteristics. To investigate personal auditory characteristics, temporal resolutions were measured by post masking line identification method. The characteristics for temporal resolutions indicated that reverberations in ears for impact noises became decreased. As the result of this measurement, subjects were classified into two groups, hereafter referred as groups "A" and "B". The subjects in group "A" had inferior ability of temporal resolution, and their post maskings for impact noises were reduced slowly. The subjects in group "B" had superior ability of it, and their post maskings were reduced extremely.
    Subjects of groups "A" and "B" were then exposed to intermittent noises with different pause durations. All subjects evaluated these noises more clearly than equivalent sound level. Equivalent sound level was the equality with the mean value of noise-energy for these intermittent noises. In case of intermittent noises with long pause durations, subjects in group "A" evaluated the noises louder than group "B". Therefore, reverberations in ears at long pause durations influenced the loudness of intermittent noises.
    Further, by the use of intermittent noises with different sounding durations, the experiment of evaluation of the loudness was carried out. Based on the result of this experiment, subjects were grouped into two groups "C" and "D". Subjects in group "C" evaluated these noises to the same value of equivalent sound level. However, subjects in group "D" evaluated these noises less than the equivalent sound level of them.
    Finally, by the use of intermittent noises with the same equivalent sound level, the experiment was carried out, in which subjects in groups "C" and "B" evaluated these noises to the same value of equivalent sound level. Subjects in groups "C" and "A" evaluated these noises louder than equivalent sound level of them. And subjects in groups D and A evaluated these noises less than equivalent sound level of them.

A study on Human Factors for Accident Prevention by Analysis of Near-Accident Reports

RR-94-5
Shinnosuke USUI

: It has been well known that collecting and analyzing case reports on near-accidents is useful for preventing industrial accidents accompanying human error. The "near-accident" is defined here as an accident which does not result in absence of the workers from their jobs but may bring about a serious one for the worse or worst case. By use of this "collecting and analyzing" methodology, 280 near-accident cases reported by workers of an electric power company were analyzed in this paper, so as to find background situations and processes affecting human error causation. Each subject who had experienced the near-accident was asked to fill a questionnaire on such items as the time of accident occurrence, weather of the day, the work being performed, details and the probable cause of the accident, and other related information.
    Main results of the analysis are summarized as followings ;
  (1) 35% of the total accident reported were on simple action slip which might not develop into a serious accident.
  (2) Near-accidents due to misconvince, or misunderstanding, by the subject were 5% of the total, and only 1% of the whole accidents were attributed to inadequate communication between the subject and its colleagues. Those ratios were rather smaller than the expected. However, it seems important for the prevention of accident to analyze these types of error precisely, because they are often liable to cause serious accidents.
  (3) The reasons why the number is small for the accidents caused by misconvince or wrong communication are ; firstly, those mental error seldom remains in the subject's memory, compared to the physical error, such as action slip, which may give the strong impression on the subject's body and remains in its memory for a long time. Secondly, because the questions were replied voluntarily by the subjects, they were more willing to report on trivial accidents experienced by themselves, than hesitating to remember other cases which might give troubles to, or might ask responsibility for, themselves, their colleagues or the company they belonged.
  (4) Each subject might have an individual concept of "near-accident" based on its career and position in the company, and the fact should have affected the number of accident reported as well as the descriptions in each report.
    In concluding, it was found important to establish a more effective methodology which could collect satisfactory information on real features of near-accident in industrial fields, so that the role of human factors on accidents could be analyzed more precisely and definitely.

Analysis of Buckling Load on the Wedge-Type Shorings --On Buckling Load of Semi-Rigid Frames without Braces--

RR-94-6
Katsutoshi OHDO, Yoshimasa KAWAJIRI and Katsunori OGAWA

: In case of concrete works, wedge-type shorings are often used and conventional shorings as frame-type shorings tend to be replaced by them. These shorings have wedge connections and the connections are semi-rigid of which flexural rigidity is between fully rigid and pinned connections. The semi-rigid connections as the wedges contribute to increasing of buckling strength of the shorings, and these strength is usually examined by the actual buckling tests because the design method of the shorings with the semi-rigid connections is not clarified yet.
    The actual tests can not be performed so many times, therefore it is difficult to examine the strength of all wedge-type shorings assembled in various conditions. So the strength of temporary structures like these shorings is usually calculated by simple models with an assumption that the semi-rigid connection is either fully rigid or pinned one.
    Recently, some collapse accidents of the wedge-type shorings happened. For the purpose of preventing these accidents, it needs to develop a new design method to calculate the strength of almost all wedge-type shorings.
    In this paper, analyses on the buckling load of the wedge-type shorings were performed by assuming flexural rigidity of the wedge connections, which had been confirmed by bending tests of the connections. Generally the semi-rigid connections like the wedge connections were considered as rotational springs which had no length, but in these analyses the wedge connections were assumed as elastic-plastic beams which had length, because the length of the connections was very large and was thought to influence on the results of analyses.
    The analyses were performed on three types of the wedge-type shorings and the results were compared with the those of the actual buckling tests. From these analyses and actual tests, the propriety of this analysis method was evaluated.
    The results are summarized as follows:
  (1) From the bending tests of the wedge connections, it was found that flexural rigidity of the wedge connections could be expressed in two lines.
  (2) The buckling loads of the wedge-type shorings in the actual tests were 2.5 times or more than the Euler's loads, in which the connections were assumed as pinned.
  (3) From the comparison between the analyzed values and the calculated values by the simple equation which was based on the energy method, the analyzed values were greatly influenced by flexural rigidity of the wedge connections and boundary conditions of the shorings.

Failure Characteristics of Sheet Pile Wall in Centrifuge Tests

RR-94-7
Yasuo TOYOSAWA, Noriyuki HORII and Satoshi TAMATE

: Accidents due to trench failure frequently occur on construction sites. One of main causes of these accidents is shallow penetration of sheet pile wall or insufficient support such as neglecting struts etc.
   To study these kinds of sheet pile wall failure characteristics and mechanisms due to excavation, centrifuge tests were undertaken on trench model and vertical cut model with installation of sheet pile walls in the preconsolidated kaolin clay.
   The progress of deformation of the sheet pile wall during excavation was analyzed from the displacement of targets set on the model measured from a sequence of photographs taken at intervals of about 0.3 seconds until the model collapsed. From the measured lateral earth pressure around sheet pile wall and the calculated strains, the development of strains such as ε1, ε3, γmax up to collapse were evaluated.
   Based on the results of centrifuge test, the following conclusions were obtained.:
  (1) The rotational failures of wall around the prop at the top were observed in case of shallow penetration sheet pile walls. On the other hand, for the unpropped sheet pile walls, toppling failures rotating around toe area observed.
  (2) For the shallow penetration sheet pile walls with struts at the top, at the onset of failure, the earth pressure were concentrated on the penetration area sheet pile walls and then it led to the base heave below the base of excavation. In this case, the active earth pressure was growing considerably larger than the designed earth pressure.
  (3) Photographic measurements showed that, prior to failure in sheet pile walls with struts, the strains were concentrated around the base of excavation in passive side and the toe of wall in active side. However, in unpropped sheet pile walls, the strains were observed around the toe of excavation ares behind the sheet pile wall and then a crack, appeared behind the top of slope at the surface.
  (4) Width of excavation and penetration depth of sheet pile wall were the major factors of sheet pile wall stability within these test condition, narrower width of excation and deeper penetration made sheet pile wall more stable.
  (5) Because of the development of strains around the area of penetration of wall the beginning of failure, it was regarded the strength of these area of the model ground effects the stability of wall.
  (6) As mentioned above, depending on the conditions of props, the different characteristics of shear strains and earth pressures around sheet pile wall were observed in centrifuge tests.

Activated Carbon Dust Explosions --Effects of Ignition Energy on Dust Explosion Parameters--

RR-94-8
Toei MATSUDA and Haruhiko ITAGAKI

: The industrial use of activated carbons has been increased for environmental aspects, but they are predicted to show a dust explosion. Activated carbon dusts are characterized by a large specific surface area from their internal porosities, and they react heterogeneously with air at first step. However, the role of the surface reaction in dust explosion mechanism has not been well understood.
    An International Standard (ISO) specifies one cubic meter explosion vessel for realistic dust explosibility determination. To improve time and cost effectiveness of the large scale tests, a 30 L spherical closed vessel has been developed and used in our laboratory for hazard assessment of combustible dusts. The ISO standard also recommends the use of a 10 kJ chemical ignitor for the determination of dust explosion severity. However, the use of such a powerful ignitor in a smaller explosion chamber leads to overdrive of the dust-air mixtures.
    This paper reports the effect of ignition energy on explosion pressures, maximum rates of pressure rise, and lower and upper explosible dust concentrations for activated carbon dust explosions in the 30 L sphere and a standard 1 m3 cylindrical vessel.
    Within the scatter in the data, the explosion pressures at a given dust concentration are almost independent of the ignition energies used. By increasing the dust concentration up to 10 kg/m3, the explosion pressure gradually decreases even to 2 bar when the 10 kJ ignitor is used in the 30 L vessel. Each ignitor distinctively provides an ignitable range of dust concentration. Dust-air mixtures of two dusts used could not ignite at any concentration using the 1.5 kJ ignitor in the 1 m3 vessel. K st values, the size-normalized maximum rate of explosion pressure, reflect the relatively weak explosion severity of the dusts used. The effect of ignition energy on the K st is evident as the values increase with increasing ignition energies, and the 30 L vessel shows the higher K st for the dusts when the same ignitor energy is applied.
    Explosion time, or combustion time from ignition to a peak pressure, has been used as the criterion for apparent limits of explosibility in our laboratory. In the 1 m3 vessel, the lower explosibility limits for two kinds of activated carbon dusts are relatively independent of ignition energies, whereas the data from 30 L show wider apparent limits of explosibity with increased ignition energies. From a comparison of these data, the lower explosion limits for two dusts used in the 1 m3 vessel with a 10 kJ ignitor appear to be comparable to that in the 30 L vessel with a 2.0 or a 1.5 kJ ignitor. Thus, the result for the activated carbon dusts tested shows that the appropriate ignition energy for lower limit determinations, which should be used in the 30 L vessel in order to obtain comparable data to those in 1 m3 vessel, seems to depend on the kind of dusts being tested.
    The upper limits for the activated carbon dust in both vessels show a large difference of more than 10 times in concentration with a 10 kJ ignitor. Although such a discrepancy would be explained by different turbulent levels in both vessels, the increase in the apparent upper limit concentrations with ignitor energy in the 30 L vessel would be promoted by the higher energy ignitor itself which preheats the excess dust particles acting as a heat sink in the denser mixtures. Nevertheless, it should be noted that the explosion pressure recorded with a 10 kJ ignitor near the upper limit in the small vessel is nearly 2 bar, as above mentioned. This value is close to the pressure criterion previously defined for the limit concentration. The upper limit of explosibility for the dust in the 30 L vessel would be then justified and therefore the 30 L vessel would be useful to reveal the potential hazard of such a dust.

Thermal Ignition of the Activated Carbon Used for Waste Treatments

RR-94-9
Haruhiko ITAGAKI and Toei MATSUDA

: Thermal ignition properties were tested for the activated carbons adsorbed with sulfur oxide (SO2), nitrogen oxides (NO2, N2O) or fluorine (F2) by using a differential scanning calorimeter (DSC) and an Accelerating Rate Calorimeter (ARC), to clear the thermal reactivity of the activated carbon used for waste treatments. The activated carbons selected in this paper are used for chemical experiments, for gas chromatography and for refinement/purification in industries.
    For DSC tests, the activated carbon that adsorbed 81 ppm nitrogen dioxide or 0.08% fluorine in air shows a dangerous thermal reactivity, while the activated carbon without adsorption of any chemical substance can hardly ignite by itself. When the concentration of nitrogen dioxide in air increases to 10%, the exothermic onset temperature drops to around 47 °C, and calorific value increases to 104 cal/g for DSC tests. For ARC tests, the sample adsorbed with 10% nitrogen dioxide in air initiates self-heating at around 30 °C, and the exothermic reaction continues up to the temperature of 150 °C intermittently, though the self-heat rate does not increase. For higher concentrations of nitrogen dioxide than 40%. the activated carbon changes to redhot, then it ignites by itself during the adsorption. Cautions should be paid in selecting the kind of activated carbon, size of adsorption column, flow rate and concentration of processing gas to prevent explosion/fire accident, when oxidative gases are treated with activated carbon.

Influence of Functional Groups on Thermal Explosion of Nitrobenzene Derivatives

RR-94-10
Yasuhiro FUJIMOTO

: Evaluation of thermal hazard of unstable chemical substances is important in the chemical industries because explosions or fires occasionally occur in chemical processes in which fine-chemicals including Pharmaceuticals or functional resins are produced. If the evaluation is possible by easy calculation, it may save chemical industries money and time.
    Thermal hazards of chemical substances must be evaluated from the two aspects : the first is sensitivity, which shows how easily a chemical may decompose or ignite, and the second is severity, which shows how much or how fast energy is liberated when a chemical decomposes. One of the typical methods for the evaluation of thermal hazards is DSC, which gives information on the sensitivity (or exothermic onset-temperature) and the severity (or decomposition heat) at the same time with easy operation.
    Prediction trial of thermal hazard data (DSC-data) from chemical structure by use of multivariate statics analysis and fuzzy reasoning was not successful perfectly in the previous work. One reason is that the mutual interaction between functional groups (for example, hydrogen bond) may decrease or increase the activation energy of the thermal decomposition.
    To obtain further understanding on thermal decomposition characteristics of unstable substances, focus is put on the influence of functional groups of nitrobenzene derivatives on thermal explosion indices. In this paper, the exothermic temperatures of the derivatives by DSC are compared with
  (1) information from NMR and IR,
  (2) decomposition process in thermal decomposition through analysis of intermediate products.
    The information from NMR or IR was not so sensitive for functional group's positions on benzene ring.
    Decomposition process was, however, quite different in ortho-and the other derivatives for some chemicals. For example, some ortho-nitroalkylbenzenes had lower exothermic onset-temperatures and decomposed via special ring formation. Thus, in some cases, decomposition process could explain the difference of exothermic onset-temperatures between ortho-and the other derivatives.

Simple Method of Test for Oxidation Capacity of Solid Chemicals

RR-94-11
Yasuhiro FUJIMOTO

: It has widely been discussed on the definition and classification of dangerous chemical substances for the purpose of their safe production, handling and transportation. From a safety point of view, it is preferable to determine by appropriate tests whether or not a chemical substance has hazardous properties such as detonability and high combustibility, or in which group the chemical is classified. Though various kinds of test method are presented so far for the evaluation of hazardous substances, it is not easy to carry out all of these tests because they require higher cost and tremendous time for evaluation. In addition, recently the ratification of the Basel Convention has forced us to control international transportation of hazardous wastes. The domestic regulation based on the Convention requires to test and evaluate if a waste is hazardous or not. It is impracticable, however, to apply those methods enforced by the regulation for testing "waste" with widely changed compositions and physical characteristics.
    In this paper, a simple test method is proposed for testing oxidation capacity of chemicals in the shape of powder or granule. The oxidation capacity is worldwidely classified as one of the hazardous properties of chemicals. The standard method for testing oxidation capacity is the Burning Test as defined, for example, in the Fire Prevention Law. In this method, each test shall be performed on the mixture of wood-powder and the chemical with total mass of 30g, thus a test necessarily gives a hreatening of burn to the examiner in case of highly oxidizing substance, as well as much amount of smoke in a test area. So the test is recommended to be done in a specified draft chamber, which is usually expensive. Thus, in developping a revised simple method, the focus has been put mainly on the amount of mixture under test. The proposed method is to determine the time for complete combustion of the test mixture on an aluminum foil by use of a controlled Bunsen instead of applying a red-hot Nichrome wire specified in the standard method. The method in this paper requires only 2g of mixture for each test, 1g of cellulose powder and Ig of substance under test. Then it gives less possibility of burn and less smoke, and needs no additional equipment or area for the test. Result of Burner Tests for more than 25 oxidizing reagents has shown that they could be classified into three groups : the first group burns completely in 5 minutes ; the second group includes such substance which never burns with flame but is carbonized completely within 5 minutes : substances carbonized only partly in 5 minutes' heating are classified into the third group. Comparison of the present results with those by the standard method has shown satisfactory correlation between their classifications of hazard.
    The proposed method for testing oxidation capacity may not be a substitute for the standard one, but is distinctly useful in pre-screening of the hazard degree of oxidation capacity of industrial wastes.

Explosion Venting into an Enclosed Vessel

RR-94-12
Haruhiko ITAGAKI and Toshihiro HAYASHI

: Explosion relief venting has been used for protecting equipments which might cause an internal explosion and for minimizing destructive effect of the explosion. In some cases, however, venting of an explosion into atmospheric air is not preferable because of harmful or poisonous materials processed in such equipments.
    In this paper, venting of gaseous explosion into an enclosed vessel was studied by explosion tests and theoretical analysis. Theoretical calculation was based on a simple mathematical model assuming ideal conditions of combustion and adiabatic expansion of gas.
    Experiments were performed with near stoichiometric propane-air mixture by use of two sets of interconnected cylindrical vessels, whose diameter/length ratios were unity. The mixture was ignited in a vessel, "ignition-side vessel", of internal diameter Dv1=200 mm, and the explosion was vented through a circular opening covered with a thin aluminum diaphragm into an larger enclosed vessel, "vented-side vessel", filled with gaseous nitrogen, diameter of which was either, Dv2=300 mm or 500 mm. Diameter of the vent opening, Do, was ranged between 6.3 mm and 160 mm for each combination of vessels. Pressure-time curves in the ignition-side and vented-side vessels were recorded in each test, and the maximum pressures attained in the vessels were plotted against the diameter of the vent opening.
    Explosion tests showed the distinct effect of the vent opening diameter on the pressure attained in the ignition-side vessel, which decreased with increase of the opening diameter up to 20 mm, but was almost constant for larger openings in both combinations of vessels. Maximum pressure attained in the vented-side vessels showed rather slight dependency on the vent opening diameter except for smaller diameters than 10 mm.
    Calculated pressures in ignition-side vessel of each setup showed satisfactory agreement with those observed in experiments, if the opening diameter was smaller than 20 mm. Predicted values of pressure were, however, far lower than experimental ones for larger openings than 20 mm, as well as for pressures in the vented-side vessels. One of the probable reasons for the difference between the calculated and the experimental ones might be attributed to combustion of unburned mixture vented into the vented-side vessel. Even when the vented-side vessel had been filled with nitrogen as an inert, combustion could not be suppressed in the vessel for larger vent openings, because of the fairly large volume of the mixture gushed into the vessel. Discussions were made to improve the mathematical model for predicting pressure behaviours in an enclosed venting system.

An Experimental Study on Amplitude of Saturable Reactor Output Current for Application to DC Ground Fault Detection

RR-94-13
Eiki YAMANO, Kenji ICHIKAWA and Tatsuo MOTOYAMA

: Detector of DC ground fault current is required today in various situations. In saturable reactor which is a main component of magnetic amplifier, DC input current changes the firing angle i.e. starting phase angle of saturation. It results in change of mean value, peak value, and pulse width of output current. The peak value is more sensitive to input DC current than the mean value.
    In this investigation, dependencies of the relation between DC input current and peak output current on various conditions were examined. The maximum gain was 2 times, and was obtained by saturable reactor made of amorphous core with 10 turns secondary winding on the conditions of 8 kHz exciting frequency and small initial saturated current. Number of primary turns and load resistance were set to 1 turn and 10 ohms, respectively.
    It has been found from the experiment that a magnitude of calculated magnetic flux for unsaturated condition relative to real saturating flux is important to set or estimate the parameters such as number of turns, exciting frequency and exciting source voltage, because these parameters are included in the calculated flux. The magnetomotive force by secondary winding should be also considered to obtain high gain in current amplitude using a saturable reactor.

Investigation of Transition of Discharge Modes in Radio Frequency Discharge Used in Plasma Processing

RR-94-14
Atsushi OHSAWA and Yasuyuki TABATA

: Capacitively coupled radio frequency (RF) discharges have been widely used in industrial applications, e.g., fabrication of microelectronic devices, gas lasers and decomposition of pollution gases, because the discharges provide higher energy electrons and ions, excited atoms and molecules and radical species, then they lead to high temperature chemistry at low gas temperature. However, the understanding of the mechanisms of discharge sustaining and chemical reaction in the process is not rich. A great deal of knowledge of the mechanisms is essential for optimized and safety control of the discharges.
    In this paper, the electron temperature, the electron density, the electron energy distribution function (EEDF), the light emission and the DC self-bias voltage in argon RF discharges for frequencies of 5, 9 and 13.56 MHz, RF voltage amplitude up to 1600 V and a pressure of 0.4 Torr have been measured to investigate the discharge transition from low RF voltage (a) to high RF voltage (γ) mode. When the transition occurs, these measured characteristics changes dramatically. The changes show the existence of two distinct discharge modes. Probe measurements showed the transition was accompanied by a sharp drop in the electron temperature, a sharp rise in the electron density and a decrease of higher energy region in the EEDF. Optical emission spectroscopy also showed the existence of two discharge modes which drastically changed in the light emission spectra and axial distribution of the light emission. In the high voltage mode, it was found that the axial distribution consisted of the light emissions caused by a direct excitation and a recombination, while the direct excitation was a main process in the low voltage mode. The DC self-bias voltage as a function of RF voltage amplitude also changed, when the sheath voltage of the driven electrode achieved a sustaining voltage for a DC normal glow discharge.
    These results emphasize that users of the discharge must pay attention to the distinct different discharge modes for an optimum and safety control of the plasma processes.

Other Publications

Kiyose District Map

Umezono 1-4-6, Kiyose,
Tokyo 204-0024 Japan
TEL: +81-42-491-4512
FAX: +81-42-491-7846

Noborito District Map

Nagao 6-21-1, Tama-Ku,
Kawasaki 214-8585 Japan
TEL: +81-44-865-6111
FAX: +81-44-865-6124