金屬膜の電氣的性質に關する豫備的實驗

(1) Cathode Sputtering。Sputtering の方法は大體Bossの方法によつた。試片は短冊形の硝子片に着膜した。唯從來の實驗と違つてゐる點は,陰極板の代りに金屬の粉末をも使つたこと,又布や紙を硝子板の代りにして,其を鍍銀して見たこと等である。(2)膜の外見。非金屬的で薄膜に於ける干渉色でない色を持つ膜,Biでは特に白色な膜,普通の金屬光澤の膜,黒味を帯びた灰色の膜,及び其等の中間の種々な膜を經驗した。非金屬的膜の發生原因は不明である。Biの白色膜發生の原因は,放電による陰極の加熱の爲に,硝子板が熱くなるのを主因と考へた。此考へは,放電時間の斷續,放電條件の調節,硝子板背面に放熱器を貼る事等によつて,裏書きされた。併し此等の膜の均整に必要な條件を充しても,猶ほ部分的に現れて來る反射能の不均等に就いては,熱電氣測定と相俟つて,陰極板に或る固有な因子のある事を知つたが,單結晶陰極板を用ゐることによつて,陰極の結晶状態のみが主因をなすものでないことを知つた。併し此の陰極板表面の影響の主因は不明である。(3)顯微鏡寫眞。反射能の小さい膜では明白な粒状組織が見られる。其上に屡々数倍の大粒が散布してゐる。此はBi, Pt, Pd, Agの皆に共通である。特にBiでは,大粒が丘の樣な形になつて存在する事を知つた。尚ほ一つの試みとして,擬格子的粒状構造の数學的分類法を考案し,此種の雜然たる組織の分類法として應用され得る事を指示した。併し,金屬膜の寫眞に就ては顯微鏡の誤差に沮まれて,未だ充分な結果は得られなかつた。又實際の場合に類似した,人工的格子組織の寫眞を作つて,同じ方法でphotometerを使つて比較する事をした。此によつて硫黄の膜が部分的には正方格子に近いと云ふ結果を得た。(4)電氣抵抗。外見上非金屬色を呈する金屬膜は異常に大きな抵抗を持つ。白色膜は金屬光澤膜に比して餘り異らない。白色及び種々の反射能の膜の比抵抗と厚さとの關係は興味ある問題であるが,併し,此れは此處で試みた樣な膜の構造の研究と相俟つて行ふべきものと考へられる。二種の異金屬の膜の二重層を作つて,其抵抗を測つた。Biが關係する膜には,著しい抵抗の特異性のある事を認めた。併し量的の結果は後日に讓る事にした。(5)熱電氣的性質。Bi膜に就ては,同一金屬の異種の膜の間に著しい熱動電力を認めた。此E.M.F.は反射能の良い膜が反射能の惡い膜に對して,hot junctionでelectro-positiveと云ふ結果に達した。放電の斷續,放電條件の加減,硝子板背面に放熱器を付けること等によつて支配した反射能の差に於て,上のE.M.F.と反射能との關係が成立つことを認めた。此と同様にPt, Pdに就ても,放電條件を變へて付けた膜の間に明瞭なE.M.F.の起ることを知つた。要するに膜の熱電氣的性質の研究には,其反射能と關係ある顯微鏡的構造の研究が必要であることは,以上の結果から明かである。又一方では,超顯微鏡的構造の差を熱電氣的性質から研究し得る曙光を認めることも出来たと考へる。又此の構造の研究は,結局は當然電氣抵抗の本質に關する研究と連關して來るであらうと考へる。終りに,此實驗は終始;寺田寅彦先生の懇篤な御指導に頼つて成つたものである。

This paper contains the reports of miscellaneous preliminary experiments which were considered necessary to be carried out before entering upon a special research on properties of sputtered metallic films as the continuation of a previous study on the allied subject. Though the results are yet incomplete in many points, they are summarised here for future references. In the experiments described in the preceding report, it was difficult to obtain a thermo-electrically uniform bismuth film. In the present experiment, however, it was found that a sufficient uniformity could be obtained by keeping the glass strip, upon which the metal is to be deposited, as near the room temperature as possible. For the control of temperature, it was found most practical to use intermittent sputtering. Besides, many other conditions such as the pressure in the discharge tube, the magnitudes of the electric current density, and cathode fall, or such an auxiliary contrivance as attaching a metallic plate as a thermal radiator on the back of the glass strip, have also some effects on the uniformity of the film obtained. The following is a brief summary of the full report. (1) Cathode sputtering. The metals used was Bi, Pt, Pd and Ag. The method of cathode sputtering was similar to that of Boss. The metals were deposited on glass plates, about 1.5 cm wide, 20 cm long, and 0.5 cm thick. Among the other details of technical interest during the course of experiment, it was found that a film with good reflecting power can be produced by rubbing graphite powder into the bismuth cathode. Though the film thus obtained was found to be probably containing the impurity of carbon, by the study of its thermoelectric property, the sputtering method may in some cases be applied with advantages for the purpose of getting a good reflecting mirror when no claim is made of the purity of metal. The trial of using metal powder as the cathode proved successful, putting selenium powder in a shallow aluminium vessel. By this method, it may be possible to get a film of mixed metals of any desired constitution, without using alloys. Sputtering on other substances such as pieces of cloth, paper, goldbeater's skin, or those painted with celluloid, or a dope, was tried with good results, except in the case of goldbeater's skin. Paper was made electrically conducting without sensibly weakening its strength against tension, but the sputtered pieces of cloth were utterly nonconducting. This conducting paper might be profitably used for constructing a balloon in the case when it is desirable to make it electrically conducting, with a small increase of weight and expense. (2) Appearance of films. Some films show non-metallic colour somewhat different from the usual interference colour of thin films. Especially in the case of bismuth films, different varieties are obtained such as milky white or dark grey films, besides those with usual metallic appearance. The origin of the films with non-metallic colour is not yet clear, but that of the white bismuth films was inferred to be due to the rise of temperature of the glass plate by the sputtering process. This inference was made from the results of sputtering with the control of temperature of the glass strip as described above. (3) Microphotograms. The microphotograms of the surface of some film with inferior reflecting power show conspicuous granular structures. Superposed on that structure there are also, frequently, a distribution of nearly spherical grains of different orders of magnitude. In bismuth films, these spherical grains are seen to be grown up to hillocks with irregular angular forms. These are shown in Plates. Considering the granular structures as consisting of a pattern statistically resembling to a plane-lattice, a mathematical analysis of the photograms was attempted as follows. On the photogram is laid a long slit narrower than the average grain size, and the number n is counted of the white spots falling in the slit. Consider n as a function of the angle of direction of the slit θ. Varying θ successively the form of the function is determined by the usual harmonic analysis. Taking photograms of the same spot of the film at various relative orientation of the specimen and microscope, we obtain different harmonic coefficients which are determined by the real film structure as well as by the optical error of the microscope. We can analyse these effect by means of the least square method, assuming that they are superposed additively. Afterwards a microphotometer was used in place of counting n by the slit. On the other hand the photometer curves obtained with the photogram of the metal film was compared with the corresponding photometer curve of an artificial regular lattice of the similar grain size. Though the results were rather inconclusive owing to the too large optical errors of the microscope, the method here initiated may have many important applications in analysing such irregular patterns of structures as shown by microphotograms of metallic alloys, photograms of clouds, etc. (4) Electric resistance. The films with non-metallic appearance have generally large resistances of different orders of magnitude. The resistance of white bismuth films is however, not much different from that of the films with usual metallic appearance. Two different kinds of metals were deposited on a glass strip one after the other, and thus a composite film could be produced. When bismuth is one of the two metals, the double layer shows generally an anomaly of electric resistance. All the quantitative relations regarding this point are reserved for the future study. (5) Thermo-electrical property. The measurement was made by heating the succcessive narrow elementary portions of the film with the radiation from a tungsten C-lamp, keeping the other parts at the room temperature. The deflection of the mirror of a sensitive moving-coil galvanometer was photographically recorded on a rotating drum covered with photographic paper. To discover the main cause of the irregular distribution of the thermo-electromotive forces observed on bismuth films, the results for a number of specimens were examined statistically. An influence proper to the nature of each cathode plate on this distribution was found to exist. A monocrystalline cathode plate was also used, but the irregular distribution could not be eliminated. The cause of the irregularity is still unknown, though it is probably connected with the irregular distribution of temperature on the glass plate caused by some unknown reason. A relative thickness measurement of the different parts of a bismuth film was made, and the thicker parts were found to be relatively electro-positive at the hot junction. Microphotograms of various parts of a bismuth film were taken, and comparing them with the thermo-electrical curve, the parts with the denser distribution of the spherical grains described above in (3) were found to be the more electro-negative at the hot junction. Depositing on different parts of the same glass strip two kinds of films of the same metal sputtered at different temperatures, of which the control was made by the method described above, it was found that the more reflecting the film the more electro-positive it is at the hot junction. By this experiment, it was found that the thickness plays no large part in the thermo-electric property. The accidentally produced thicker parts were always less reflective, and by this reason only, the apparent relation between the thickness and E.M.F. had been obtained. In this way, films with various thermo-electromotive force can be produced from the same kind of metal, except Ag with which no example of this difference of thermo-electromotive force has yet been obtained. Bismuth films show the greatest difference of this E.M.F. with a different order of magnitude compared with the other noble metals examined, i.e. Pt and Pd. The differences of E.M.F. for these noble metals are, however, not very small, so that it is easily measured by a sensitive moving-coil galvanometer, with the tungsten lamp heating. By these preliminary experiments, it is confirmed that the thermo-electric property of a metal film has an intimate relation with its microscopic structure which is connected with its reflection coefficient. Interchanging the point of view, the microscopic structures of metal films can be studied in some measure from the difference in their thermo-electric property. These experiments were carried out under the guidance of Prof. T. Terada.

資料番号: SA4415977000