STUDIES OF SHORT-WAVELENGTH ULTRAVIOLET RADIATION FROM THE SUN

S. L. Mandelstam and A. I. Efremov

Studies of recent years have greatly expanded our knowledge of the Sun’s radiation.

The use of high-altitude rockets has made it possible to detect and study short-wavelength ultraviolet radiation, strongly absorbed by the atmosphere and therefore not reaching the Earth’s surface, up to the soft X-ray region. New possibilities for these investigations will be provided by the use of artificial Earth satellites, which make possible prolonged measurements inaccessible when rockets are used.

Absorption in the Earth’s atmosphere practically cuts off all solar radiation with wavelengths shorter than 2900 Å. Meanwhile, this short-wavelength part of the solar spectrum is of exceptionally great interest. Into this region, down to 1700–1600 Å, extends the radiation of important portions of the photospheric spectrum; in the region 1600–900 Å the principal radiation of the chromosphere is concentrated, and in the region shorter than 900 Å (down to several angstroms) lies the radiation of the corona. The chromosphere and corona emit very little in the visible and near-ultraviolet region of the spectrum accessible to terrestrial observations; therefore detailed knowledge of these envelopes of the Sun can be obtained only by investigating their short-wavelength radiation. At the same time, the study of this short-wavelength radiation is also extremely important for many problems in the physics of the Earth’s atmosphere, since this radiation ionizes air molecules and is responsible for the formation of ionospheric layers in the Earth’s atmosphere.

In the present review, new results of experimental and theoretical work on the study of the Sun’s short-wavelength radiation, published since the first review on this subject¹, are briefly set forth; experiments that are expected to be carried out with the aid of artificial Earth satellites are also described.

§ 1. RADIATION OF THE PHOTOSPHERE

The principal report published in the literature in recent years is the paper by Klearman², in which detailed results are given for the processing of spectrograms obtained earlier by Hopfield and Klearman³. The paper gives tables for all absorption lines that could be measured in the spectrum. The spectrum studied covers the region 2285–3000 Å and contains 248 lines. In the table, for each line, are given the measured wavelength (the accuracy of the measurements is estimated at 0.2 Å), a visual estimate of intensity, the elements to which the given line may belong, and the laboratory wavelength of the line. It proved possible to identify about 90% of all the lines. The principal results of this identification by elements are as follows:

Bi — the presence of this element on the Sun was established only from molecular bands. Its most sensitive lines \((\lambda = 2496\ \text{Å}\)

and \(\lambda = 2497\) Å) fall in a region of the spectrum containing many strong iron lines and could not be detected.

C I—the only intense line in the region under consideration, \(\lambda = 2479\) Å, is blended with Fe lines and could not be detected.

O I—a multiplet consisting of three lines is attributed to this element. These lines, however, are overlapped by Fe and V lines, as a result of which the identification is not considered reliable.

Na I—two lines belonging to the principal series have been identified.

Mg I—a whole series of lines has been identified.

Mg II—there is a strong resonance doublet, revealing the presence of emission lines at the center of absorption lines.

Al I—two doublets have been identified.

Al II—one line has been identified.

Si I—seven intense multiplets have been identified.

P I—five intense lines have been identified.

Elements of the long period—only a few lines have been reliably identified.

Fe I, Fe II—all principal lines falling in the spectral region studied have been identified; they constitute the overwhelming majority of all observed lines.

Molecular bands—some lines could have been interpreted as bands of certain diatomic molecules; however, the absence of the other members of the corresponding systems did not in any case permit this identification to be accepted.

Fig. 1. Doublet Mg II \(\lambda\lambda\) 2795.523 Å (A) and 2802.698 Å (B).

Thus, no essentially new results were obtained in the work. To obtain new data, photographs of the spectrum with considerably greater dispersion and resolution are necessary.

The paper gives a very interesting photograph of the Mg II doublet \(3s^2S - 3p^2P^0\) (Fig. 1). The spectrogram was obtained with an Aerobee 3/IX 1952 rocket from an altitude of 77–85 km, using a spectrograph with a diffraction grating of radius 40 cm, 600 lines/mm. The lines have an emission center against a background of broad absorption lines. The line width, of the order of 0.7–0.8 Å, is, in the authors’ opinion, the instrumental width.

In paper \(^{5}\), spectra of the Sun obtained at different altitudes between 35 and 70 km are used to study the distribution of concentra-

tion of ozone with height. An interesting design of a doubled spectrograph with a very wide field of view was used.

A theoretical consideration of radiation in the region 3000–2000 Å was carried out by de Jager^6. This radiation belongs to the photosphere and to the transition region between the photosphere and the chromosphere, and was calculated by de Jager on the basis of a certain mean model of the structure of this region. In Fig. 2, taken from this work, a comparison of theoretical and experimental results is presented. The experimental results are taken from work^7. The depression in the region 2800–2850 Å is due to the intense Mg II doublet, and the depression in the region 2500 Å is due to the limit of the Mg I series. For the region \(\lambda > 2600\) Å, fairly good agreement is observed between the theoretical and experimental data; for the shorter-wavelength region of the spectrum the discrepancies are rather significant. Thus, for 2300 Å the theoretical value of the radiation intensity corresponds to a radiation temperature of 4500°, while the experimentally found value is 5000°. The experimental value exceeds the theoretical one by approximately a factor of 5. A possible explanation of these discrepancies lies in the temperature inhomogeneity of the photosphere. Temperature inhomogeneities of the order of \(\pm 500^\circ\) can increase the radiation in the region \(\lambda = 2200\) Å by a factor of five.

Fig. 2. Energy distribution in the region 3000–2000 Å: — experiment; —·— theoretical values.

§ 2. RADIATION OF THE CHROMOSPHERE

Studies of the short-wavelength radiation of the chromosphere were carried out both with spectral decomposition of the light by means of spectrographs and with photoelectric receivers isolating a narrow spectral region.

The use of tracking systems^8, which direct the optical axis of the spectrograph toward the Sun during the greater part of the rocket’s time at the high-altitude portion of its trajectory, made it possible to obtain spectra with long exposures and to advance considerably farther into the short-wavelength region of the spectrum in comparison with the first photographs of the solar spectrum.

The first spectrogram of this kind was obtained in the flight of an Aerobee-type rocket on December 12, 1952^9. A more detailed description of this investigation is given in the work of Rense^10. A spectrograph with a diffraction grating of 600 lines/mm in grazing incidence was used. The radius of curvature of the grating was 49.80 cm, and the angle of incidence 85°. The spectrograph was pointed at the Sun by means of a two-axis tracking system; the effective exposure-

tion was 28 sec, the mean ascent height of the rocket during the exposure was 81 km.

The spectrum obtained had, in the region 5000—2000 Å, the usual appearance—numerous absorption lines against the background of a continuous spectrum. The darkening in the region of the spectrum shorter than 2000 Å was due mainly to scattering of light in the spectrograph; however, absorption lines could be traced to 1800 Å, and the continuous spectrum to 1700 Å. Shortward of this region only one line was detected; measurements of its wavelength gave the value \(\lambda = 1215.5 \pm 1.0\) Å, and the line was identified as the \(L_\alpha\) head line of the Lyman series of hydrogen. The line has a narrow emission center and asymmetric emission wings; the total width of the line is about 5 Å, and the radiation flux in the line is \(0.5\ \mathrm{erg}/\mathrm{cm}^2\ \mathrm{sec}\) at the boundary of the Earth’s atmosphere.

Fig. 3. Photograph of the solar disk:
a) in \(L_\alpha\) rays (negative),
b) in H\(_\alpha\) rays (positive),
c) in Ca II rays (positive).

In work \(^{11}\) spectra were obtained during the flight of a rocket on February 2, 1954, at heights between 80 and 104 km; the instrument was pointed at the Sun by means of the following system.

To reduce scattered light, a plastic replica grating was used, and the Sun was focused on the slit by means of a quartz mirror, whose reflectivity increases for the spectral region shorter than 2000 Å and reaches a maximum of about 20% for \(\lambda \sim 1200\) Å. An intense \(L_\alpha\) line and traces of the continuous spectrum in the region 1100—1300 Å, corresponding to the “window” in the absorption of \(O_2\), were observed. The continuous spectrum in the remaining regions was absorbed by \(O_2\) molecules located above the rocket. The radiation flux in the \(L_\alpha\) line was \(0.3—0.6\ \mathrm{erg}/\mathrm{cm}^2\ \mathrm{sec}\), and the half-width of the line was \(< 0.3\) Å.

In the work of Miller, Mercure, and Rense \(^{12}\), a spectrograph with a diffraction grating of radius 39 cm, 600 lines/mm, was used. The angle of incidence on the grating was 49.5° and was chosen from the calculation of eliminating astigmatism for the \(L_\alpha\) line. Three photographs were made at heights (the exposure duration is indicated in parentheses) of 89 km (16.1 sec), 105 km (7.5 sec), 122 km (31.4 sec). The obtained values of the radiation intensities and half-widths of the \(L_\alpha\) line are given in Table I. The elimination of astigmatism for the \(L_\alpha\) line made it possible to obtain data on the variation of the intensities of this line over the solar disk, important for constructing a model of the chromosphere.

In work \(^{13}\), the solar disk was photographed in \(L_\alpha\) rays with the aid of a monochromatic camera. The optical system

consisting of two 15° prisms and a LiF lens. The resolving power of the camera was about \(1/20\) of the solar diameter, and the width of the spectral region was about 40 Å. The diameter of the image was 2.8 mm. The camera was pointed at the Sun with the aid of a tracking system. Oscillations along the axis during the 2-sec exposure did not exceed two minutes of arc. The photographs were taken from an altitude of 148 km.

Figure 3 shows one of the photographs obtained (negative); for comparison, photographs of the solar disk in the rays of \(H_\alpha\) (6563 Å) and CaII (3934 Å) (positives), obtained on the same day (May 8, 1956), are given. There is a strict correlation between the regions of enhancement of \(L_\alpha\) and the active regions of the Sun.

Table I

The first results of measurements of the spectral region near \(L_\alpha\) with the use of photon counters, as well as with the use of special phosphors, were described in \(^1\). Further results of measurements with the aid of photon counters were obtained in work \(^ {14}\).

The sensitivity region of the counters covered the interval 1100–1300 Å. As in the preceding experiments, the radiation was detected at an altitude of \(74 \pm 2\) km; the greatest flight altitude was 128 km. The radiation intensity, extrapolated to the boundary of the Earth’s atmosphere, was \(0.10 \pm 0.02\) erg/cm\(^2\) sec. The graph giving the dependence of the radiation intensity on altitude has a straight-line form. Since near \(L_\alpha\) there is a narrow “window” in the absorption of \(O_2\), about 1 Å wide, at whose edges the absorption increases by 200%, the linear course of intensity with altitude indicates that the width of the \(L_\alpha\) line is less than 1 Å. This agrees well with the later photographic measurements given above. According to these measurements, the intensity of the continuous spectrum in the \(L_\alpha\) region has a value of 0.01 erg/cm\(^2\) sec over a region of 100 Å.

The most recent photoelectric measurements were carried out by Byram, Chubb, Friedman, and Kupperian in 1956 \(^ {15}\). The radiation receiver was an ionization chamber filled with NO and closed by a LiF window. The transmission cutoff of LiF (about 1100 Å) determined the short-wavelength boundary of the recorded region, and the ionization threshold of NO at 1340 Å determined the long-wavelength boundary. Thus, the sensitivity region of the receiver covered the interval 1100–1340 Å. In accordance with the photographic measurements, \(L_\alpha\) accounts for 95% of the total flux in this region. The measurements gave energy values exceeding 2 erg/cm\(^2\) sec at the boundary of the Earth’s atmosphere.

Comparison of all the measurement results presented above gives, for the intensity of the \(L_\alpha\) line, extraordinarily differing values, lying between 0.1 and 10 erg/cm\(^2\) sec \(^ {23}\). Undoubtedly, part of the values obtained is burdened with certain errors caused by measurement errors and by errors in extrapolating the measured values to the boundary of the Earth’s atmosphere. Apparently, however, real strong variations in the intensities of the \(L_\alpha\) line in different experiments also occur. In all the measurements described, no ...

some increased activity of the Sun. This leads one to suppose that strong variations in the intensity \(L_\alpha\) may also occur under conditions of a relatively quiet Sun (see § 4).

Fig. 4. Observed and calculated values of the energy of the continuous spectrum of the chromosphere.

A theoretical study of the radiation of the chromosphere after the work of Shklovsky considered in \({}^{1}\) was carried out by de Jager \({}^{6}\). De Jager separately considers the continuous radiation of the chromosphere and the line \(L_\alpha\). Assuming that the radiation of the continuous spectrum can be approximated by the radiation of a gray body, de Jager compares the experimental results given in \({}^{1}\) with calculation data for different values of the temperature. This comparison is shown in Fig. 4.

In comparing the theoretical and experimental data for \(L_\alpha\), de Jager starts from the experimental value \(0.10\ \mathrm{erg}/\mathrm{cm}^{2}\ \mathrm{sec}\) and from a line width of less than 1 Å; this corresponds to a radiation temperature of \(\sim 6400^\circ\), which is considerably lower than the temperature of the chromosphere \((\sim 30\,000^\circ)\). Hence the conclusion is drawn that the radiation in the \(L_\alpha\) line is determined not by the local value of the electron temperature, but is caused by scattering.

Atey and Thomas, who also consider, on the basis of the data for \(L_\alpha\), various models of the structure of the chromosphere \({}^{17}\), arrive at the same conclusion. The shape and intensity of the \(L_\alpha\) line under the conditions of a strong flare, and also for the quiet chromosphere, were calculated by E. V. Mustel and A. B. Severny \({}^{24}\).

§ 3. RADIATION OF THE CORONA

The radiation of the corona, both in earlier and in subsequent works, was investigated with the aid of photoelectric receivers—photon counters—with narrow regions of the spectrum selected by filters.

The most recent results of the investigations are set forth in the work of Byram, Chubb, and Friedman \({}^{18}\). The photon counter was a cylindrical shell of chromonickel steel, serving as the cathode, along whose axis a tungsten filament—the anode—was stretched. The opening in the shell was covered by thin metallic and organic films, serving as filters that selected the corresponding region of the spectrum. The counters were filled with a mixture of a noble gas with a quenching molecular gas. The pulses given by the counters were integrated with a time constant of about 2 sec and transmitted to the Earth by means of a telemetry system.

The most reliable results, in the opinion of the authors of the investigation, were obtained in two flights of Aerobee NRL rockets in November and December 1953. In each flight a group of four counters was used, designed to record different regions of the spectrum and different levels of radiation intensity. The counters were oriented in such a way that they gave noticeable readings only when, during the rotation of the rocket, they were directed at the Sun.

One of the counters was covered with a beryllium window 8 mm in diameter and 0.125 mm thick and was designed to register radiation shorter than 8 Å; this counter gave no counts. The spectral sensitivities of the remaining counters are shown in Figs. 5 and 6. These curves were obtained by calculating the transmission of the windows and the absorption of radiation by the counter gas, on the assumption that only the gas absorbs and that each absorbed photon produces a count (both of these assumptions were checked experimentally).

Fig. 5. Spectral sensitivity of the counter with an Al filter. Al—1.7 mg/cm², filling—9.4 mm, ethyl formate and 625 mm Ne, absorbing layer—2 cm.

Fig. 6. Spectral sensitivities of counters with filters made of organic films \([(C_4H_8O_4)_x]\). — Mylar 0.73 mg/cm², — — glyptal 0.18 mg/cm², filling—6.4 mm, ethyl formate and 750 mm Ne, absorbing layer—2 cm.

Figure 7 shows the results of measurements with all three counters during the flight of rocket NRL-16. Along the abscissa at the top is plotted the altitude of the rocket in km, and at the bottom—the mass of air above the rocket; along the ordinate is plotted the counting rate (in pulses per second).

Table II

Table II gives the maximum values of the counting rate obtained, corrected for the dead time of the counters, which is due to their finite resolving power, and for the deviation of the normals to the counter windows from the direction toward the Sun (the position of the rocket relative to the Sun was determined with the aid of photocells). In order to pass from these data to the radiation intensity, it is necessary to know the law of distribution of the radiation intensity with wavelength. This distribution was assumed to correspond to black-body radiation, and the calculations were carried out for three values of the temperature: \(7 \cdot 10^5\), \(1 \cdot 10^6\), and \(2 \cdot 10^6\) K. The results of these

calculations are given in Table III. The fifth row of the table gives the radiation intensity in the narrow region of the spectral sensitivity of the receiver, and the sixth row gives the total radiation energy over the entire investigated X-ray region. These latter data show that a temperature of \(7\cdot 10^5\,^\circ\mathrm{K}\), corresponding to a total radiation flux of \(\sim 0.1\ \mathrm{erg}/\mathrm{cm}^2\,\mathrm{s}\), best satisfies the measurements with all three types of counters. The emissivity of the Sun in this case proves to be \(4\cdot 10^{-16}\). If the temperature is taken to be \(10^6\,^\circ\mathrm{K}\),

Fig. 7. Count rate as a function of the altitude of the Aerobee NRL-16 rocket.
\(\triangle\) ascent, \(\blacktriangle\) descent — Al; \(\circ\) ascent, \(\bullet\) descent — Mylar; \(\square\) ascent, \(\blacksquare\) descent — Glyptal (ordinates reduced by a factor of 2).

then the value of the total energy, calculated from the measurements with the counter with the Al window, differs by a factor of 10 from the value of the energy calculated from the data of the other two counters. This shows how sensitive the results of the calculation of the total energy are to the adopted value of the temperature, whereas the magnitude of the energy falling within a narrow spectral interval depends only weakly on the temperature. A temperature of \(7\cdot 10^5\,^\circ\mathrm{K}\) agrees well with the temperature \(8\cdot 10^5\,^\circ\mathrm{K}\) obtained for the inner corona from radio measurements and, apparently, corresponds to a “quiet” state of the corona. The flights of the Aerobee-14 and Aerobee-16 rockets took place under conditions when solar activity was minimal.

The radiation in the spectral region near \(50\ \text{\AA}\) is comparatively stable in intensity. The situation is different for radiation in the region of the coronal spectrum \(\sim 20\ \text{\AA}\). As a comparison of the data obtained in all rocket flights shows, the intensity of this radiation is subject to strong

…radiation. I. S. Shklovskii^16, on the basis of an analysis of the visible spectrum of the corona, showed the presence in the solar corona of regions with a higher temperature, in which coronal lines are excited that require large energies for their excitation. There is every reason to ascribe to these regions an enhanced emission in the 20 Å region.

Table III

Table IV gives the results of measurements obtained during six rocket flights. In line 5 are given the values of the radiation intensity for the entire region 0—20 Å, calculated on the assumption \(T = 10^6\ ^\circ\mathrm{K}\) and \(T = 2 \cdot 10^6\ ^\circ\mathrm{K}\). If a lower temperature value is adopted, for example \(7 \cdot 10^5\ ^\circ\mathrm{K}\), then the total intensity calculated on the basis of measurements with counters having Be windows proves to be excessively large.

In line 6 are given the intensities of some coronal lines on the basis of ground-based observations with a coronagraph. A correlation is observed between the intensity of radiation in the 0—20 Å region and the intensity of the Fe XIV and Ni XV lines. Whereas the intensity of the Fe X line is approximately the same in all cases, the intensity of the Fe XIV and Ni XV lines, corresponding to a higher excitation energy, is greater during the flights of the Aerobee-9, Aerobee-10, and Viking-9 rockets, in which a considerably greater intensity of radiation in the 0—20 Å region was observed.

The radiation of the corona was also considered theoretically by Elwert^19 and de Jager^6.

The continuous radiation of the corona cannot be directly identified with black-body radiation, since the corona is optically transparent. Elwert carried out calculations for radiation caused by photorecombination of electrons and ions and by free-free transitions of electrons in the field of ions, for two temperature values: \(7 \cdot 10^5\ ^\circ\mathrm{K}\) and

Table IV

$1\cdot10^6\ ^\circ\mathrm{K}$. The total continuous radiation of the corona for $T=10^6\ ^\circ\mathrm{K}$ is $10^3$ erg/cm² sec; the spectral distribution of the radiation is shown in Fig. 8. In Fig. 9 the radiation of the continuous spectrum of the corona for $T=10^6\ ^\circ\mathrm{K}$ is compared with black-body radiation at the same temperature. The two distributions turn out to be extremely similar, and thus the spectral distribution of the continuous radiation of the corona may be approximated by a black-body distribution; the absolute magnitude of the radiation corresponds to an optical thickness of the corona $\sim 4\cdot10^{-17}$. In this, account has been taken of the presence of density inhomogeneities $Q=2$, where

\[ Q=\frac{\overline{N_e^2}}{\bar N_e^2}. \]

Of considerable interest is the calculation of the line spectrum of the corona in the region under consideration. As is known, in the spectrum of the corona in the visible region about 20 emission spectral lines are observed. These lines are visible only in the spectrum of the corona; they are absent in the spectrum of the photosphere and chromosphere; they have also not been observed in any terrestrial source. For many decades these lines could not be identified, and their nature remained completely mysterious. Numerous hypotheses were put forward concerning their origin—from the hypothetical element “coronium” to atomic combination scattering. In 1941—1942 Edlén showed that all coronal lines can be identified as forbidden magnetic-dipole

transitions in the spectra of highly ionized atoms of iron, nickel, and other elements (Fe X, XI, XIII, XIV; Ni XII, XIII, XV, XVI; Ca XI, XIII, XV; Ar X). All these lines correspond to transitions between components of the fine structure of metastable levels, and the greater part of these identifications rests on extrapolation of laboratory values of level energies. The absence in the visible spectrum of coronal lines of a series of atoms in intermediate ionization states, for example Fe XII, Ni XIV, as well as of lines corresponding to ions of other elements, is explained by individual peculiarities in the structure of the energy states of these ions.

Thus, observations of the visible line spectrum of the corona give a very incomplete and, to some extent, insufficiently reliable idea of the coronal spectrum. All the most characteristic resonance lines of highly ionized atoms, whose presence may be expected in the corona, lie in the short-wavelength region of the spectrum under consideration. Investigation of the line spectrum of the corona in this region therefore acquires primary importance.

Fig. 8. Continuous spectrum of the solar corona. The intensity \(I_\lambda/Q\) per \(1\ \mathrm{cm^2\,sec}\), falling on the interval \(\Delta\lambda = 1\ \mathrm{cm}\), at the boundary of the earth’s atmosphere: — for \(T = 7\cdot 10^5\ ^\circ\mathrm{K}\), — for \(T = 10^6\ ^\circ\mathrm{K}\). For comparison, the radiation of H recombination is shown: — — — for \(T = 7\cdot 10^5\ ^\circ\mathrm{K}\), — — for \(T = 10^6\ ^\circ\mathrm{K}\), and the radiation of the photosphere at \(T = 5700^\circ\).

Proceeding from the probable chemical composition of the corona, Elwert determined the ionization state of atoms of elements in which they can exist under coronal conditions. Considering these ions as hydrogen-like, Elwert then calculated approximate values of the wavelengths of their resonance lines. These data were compared with experimental values of wavelengths available for certain ions, and corrections were introduced into the theoretical values. All these data are given in Table V; for further calculations the wavelength values are given in the last column.

Fig. 9. Comparison of the spectral distribution of coronal radiation (—) and of a black body (...) for \(T = 10^6\ ^\circ\mathrm{K}\).

Further, Elwert calculated the intensities of these lines. Considering the excitation of ions and the elementary processes leading to the destruction of excited ions, Elwert comes to the conclusion that the main excitation process is excitation of the ion from

Table V

Continuation of Table V

of the normal state by collisions with electrons; owing to the large difference in energies between the resonant level and the following excited level, practically only the resonant level is excited. The formation of excited ions through the recombination of ions into the corresponding state plays a considerably smaller role. Further calculations show that, for radiation in the coronal lines, the corona is not transparent. Assuming that the lines have a Doppler shape, caused by the thermal and turbulent motion of the ions, with a width of the order of \(10^{-2}\) Å, Elwert calculated the line intensities for three temperature values. The results of the calculations are given in Fig. 10.

Examination of these data shows that line radiation plays the principal role in the radiation of the corona; its intensity is \(2.5 \cdot 10^{+3} Q'\ \mathrm{erg}/\mathrm{cm}^{2}\ \mathrm{sec}\), which corresponds to \(6 \cdot 10^{-2} Q'\ \mathrm{erg}/\mathrm{cm}^{2}\ \mathrm{sec}\) at the boundary of the Earth’s atmosphere (\(Q'\) is an inhomogeneity factor, which may have a value \(\sim 2\)). The total radiation flux in this case proves to depend only weakly on the adopted value of the temperature; the maximum of the radiation, however, for a temperature of \(10^{6}\ ^\circ\mathrm{K}\) is shifted considerably into the short-wavelength part of the spectrum in comparison with a temperature of \(6 \cdot 10^{5}\ ^\circ\mathrm{K}\).

Figure 11 shows the combined line and continuous radiation of the corona, summed over discrete wavelength intervals (10, 20, 50, and 100 Å).\(^{6}\)

§ 4. VARIATION OF THE RADIATION AND EXPERIMENTS WITH ARTIFICIAL EARTH SATELLITES

The results set forth in the preceding sections indicate the existence of very considerable variations in the intensity both of the \(L_{\alpha}\) line emitted by the chromosphere and of the X-ray region of the spectrum emitted by the corona. If the absolute value of the radiation intensity of the chromosphere and corona as a whole is satisfactorily described, in order of magnitude, by the theory, then the indicated variations still remain outside the framework of the theory.

There is no doubt that these variations are connected with physical processes occurring in the chromosphere and corona of the Sun. In turn, the variations

in the intensity of the Sun’s short-wave ultraviolet ionizing radiation cause strong disturbances in the Earth’s ionosphere, manifested in disruptions of radio communication, magnetic storms, etc.

Systematic investigation of variations in the intensity of the Sun’s short-wave ultraviolet radiation is therefore of primary importance.

Recently the first direct attempt was made to relate changes in the intensity of the \(L_{\alpha}\) line and of X-ray radiation to solar flares responsible for disturbances in the ionosphere leading to “fading” in radio communication[^20].

Flares last for a very short time, on the order of several minutes; in order to ensure measurements at the moment of flares, the following procedure was adopted. The radiation receivers were placed in small rockets, which were attached to balloons. The balloons rose to an altitude of 24 km, and the rockets were launched by radio at the moment when flares appeared on the Sun. The investigations were carried out from an expeditionary vessel for two weeks, beginning on July 16, 1956, in the region of the Pacific

Fig. 10. Intensities of lines in the coronal spectrum.
a) \(T = 6 \cdot 10^{5}\ ^\circ\mathrm{K}\), b) \(T = 7 \cdot 10^{5}\ ^\circ\mathrm{K}\), c) \(T = 1 \cdot 10^{6}\ ^\circ\mathrm{K}\); the intensity of the continuous spectrum for the 50 Å region and He is indicated by circles.

Fig. 11. Line and continuous (hatched) radiation of the corona, summed over discrete wavelength intervals (10, 20, 50, and 100 Å).

Ocean, at a distance of 200–400 miles southwest of San Diego. The balloons, about 21 m in diameter, were filled with helium and lifted a four-meter rocket equipped with ultraviolet-radiation receivers and a transmitting telemetry system.

As radiation receivers, ionization chambers and photon counters were used; the \(L_\alpha\) radiation, the region \(1\text{–}10\ \text{\AA}\), and the region of hard X-rays \(0.05\text{–}1\ \text{\AA}\) were recorded.

One sphere was launched daily to an altitude of 24 km. At the moment when the “fading” appeared, a signal was sent from the ship, which activated the rocket motor and the scientific apparatus. Within \(1^{1/2}\text{–}2\) minutes the rocket reached an altitude of 50–70 miles above the Earth, and the signals from the radiation receivers were transmitted to the ship. A second method for establishing the moment of rocket launch was also prepared: an optical telescope was equipped with a violet filter isolating the \(K\)-lines of Ca II, and with a photoelectric receiver, whose signal at the moment of strengthening of the calcium lines controlled the launch of the rocket. Cloudy weather, however, prevented the use of this method.

During 11 days of observations, one flare of class 1 and two of class 2 were observed.

The latter two flares, however, were not used. The class 1 flare was too weak to cause a “fading,” and was detected with the aid of a spectrograph at the observatory in Climax. A message about this flare, quickly sent by radio to the ship, made it possible to launch the rocket while the flare was still continuing. A considerable intensity of radiation in the region of \(3\ \text{\AA}\) was detected, corresponding to a temperature of about \(10^7\ ^\circ\mathrm{K}\). The intensity of the \(L_\alpha\) line remained normal, which may be explained by the fact that the rocket reached the required altitude only after the flare had passed through maximum.

The International Geophysical Year coincides with the period of the 11-year maximum of solar activity, which creates very favorable conditions for investigating variations in the intensity of the short-wavelength radiation of the Sun. Artificial Earth satellites can evidently ensure the carrying out of these investigations, making it possible to perform reliable measurements of both long-term and rapid variations in the intensity of solar radiation.

One of the first experiments planned in the USA is the measurement of the intensity of the \(L_\alpha\) line and of variations in its intensity associated with flares on the Sun\(^{21, 22}\); it is proposed to use an ionization chamber with a LiF window, filled with \(NO^{15}\).

The first stage of the investigations planned in the USSR consists in measuring solar radiation in the X-ray region of the spectrum. The isolation of individual portions of the spectrum in this region will be carried out with the aid of filters. It is proposed to isolate the following regions: from 3–5 to \(8\ \text{\AA}\); from 8 to \(22\ \text{\AA}\); from 44 to \(120\ \text{\AA}\). As was indicated above, the first region of the spectrum corresponds to especially “hot” regions of the corona, the region \(8\text{–}22\ \text{\AA}\) corresponds to the radiation of the regions of condensation of the corona, and the region \(44\text{–}120\ \text{\AA}\) corresponds to the radiation of the main part of the corona.

Since one of the main tasks of the experiment is the investigation of long-term and short-period variations in the ratio of the radiation fluxes in the indicated regions, it is advisable to employ a method of measuring the radiation fluxes by means of a single receiver. This makes it possible to avoid errors associated with different changes in the sensitivity of the individual receivers during their prolonged stay on the satellite. In front of the radiation receiver, various filters isolating the required regions of the spectrum are installed alternately.

The block diagram of the instrument for recording the X-ray portion of the solar spectrum is shown in Fig. 12. In front of the X-ray radiation receiver 1 is placed a disk with a set of various filters 2. By means of a stepping mechanism 3, alternate rearrangement is carried out

of various filters. Signals from receiver 1 are fed to the radio-technical counting-and-integrating circuit 4, whose output is connected to the transmitting radio-telemetric system. The operation of the stepping mechanism is controlled by means of a relaxation generator 5. For timely switching-on of the instrument there is an automatic device 6, which is controlled by means of two photoresistors 7.

The instrument contains three receivers 1, spaced around the perimeter at angles of 120°, in order to increase the probability that solar radiation will fall on the photocathode of a receiver for different orientations of the satellite

Fig. 12. Block diagram of the instrument for recording the X-ray portion of the solar spectrum.

with respect to the Sun. The radiation receivers are secondary-electron multipliers made of beryllium bronze. The photocathode of these receivers is insensitive to the visible region of the solar spectrum (the red boundary of the photoelectric effect lies in the region 2300–2500 Å). The filters placed in front of the receivers are films of beryllium, aluminum, and polyethylene of various thicknesses. By subtracting the transmission of some films from the transmission of others, different spectral regions of the Sun’s radiation can be isolated. Figure 13 presents calculated transmission curves for various filters.

Region A is obtained by passing the radiation under study through beryllium foil of thickness 200 μ. Region B is obtained by subtracting, from the transmission of an aluminum film of thickness 5 μ, the transmission of beryllium foil of thickness 200 μ. Region C is obtained by subtracting, from the transmission of a polyethylene film of thickness 3 μ, the transmission of a beryllium film of thickness 4 μ.

In principle, isolation of the \(L_\alpha\) line is possible with the aid of two LiF filters with slightly shifted short-wavelength transmission limits, placed in front of a receiver whose photocathode has reduced sensitivity to the region of the spectrum longer than 1700 Å.

Sequential rearrangement of the filters is carried out at a filter-changing frequency of 2 filters per second.

In the disk of the stepping mechanism, in addition to the filters, there are free apertures through which the radiation falls directly on the photocathode. From the signals from the receiver at the moment when a free aperture is placed before it, the necessary correction for the angle of incidence of the radiation can be made.

on the filters, which changes in connection with the rotation of the satellite, and also to make corrections for changes in the sensitivity of the receiver.

The signals from the receiver, in the form of voltage pulses whose number per second is proportional to the intensity of the radiation incident on the photocathode, are fed to the input of a counting-integrating radio-engineering circuit. It consists of a pulse amplifier, a pulse converter, and three integrating circuits. At the output of these circuits constant voltages are produced, proportional to the pulse counting rate. The three outputs of these circuits correspond to three counting rates: 500 pulses/sec, 5000 pulses/sec, and 50,000 pulses/sec. The outputs of the integrating circuits are connected to the radiotelemetric system, in which these signals enter

Fig. 13. Calculated transmission curves of the filters.

a storage device and are transmitted to Earth. The telemetric system is switched on according to a definite time schedule.

The use of the pulse method of recording radiation, despite its complexity in comparison with the method of recording by direct current, is due to its greater sensitivity and to the small dependence of the measurement results on the instability of the power supplies and of the amplifying characteristics of the receiver.

The instrument is put into operation automatically at the moment when solar radiation falls on the photocathode of one of the receivers and when the radiotelemetric system is operating simultaneously. The switching-on is controlled by means of photoresistors, light falling on them simultaneously with its falling on the photocathode. Two photoresistors are associated with each receiver: one photoresistor has a field of view of \(120^\circ\) and switches on the filament circuits of the instrument; the second has a field of view of \(60^\circ\) and switches on the anode circuits of the instrument. The instrument is switched on by means of the photoresistors only when a signal for operation of the radiotelemetric system is supplied from the satellite’s special program device.

The next stage of investigations with the aid of satellites should be the obtaining of line spectra of the chromosphere and corona of the Sun. For this purpose two small spectrographs will be used, one of which covers the spectral region from 30 Å to 300 Å, and the other from 600 Å to 1500 Å. It is proposed to use two methods of recording: photographic and photoelectric.

It is also proposed to photograph the solar corona in X-rays by means of a special coronagraph.

REFERENCES

1. S. L. Mandelstam, UFN 46, 145 (1952).
2. H. Clearman, Astrophys. J. 117, 29 (1953).
3. J. Hopfield and H. Clearman, Phys. Rev. 73, 877 (1948).

1. F. Jonson, J. Purcell, R. Tousey, N. Wilson, Astrophys. J. 117, 238 (1953).

2. F. Jonson, J. Purcell, R. Tousey, K. Watanabe, J. Geophys. Res. 57, 157 (1952).

3. C. De Jager, Ann. de Geophys. 11, 1 (1955).

4. F. Jonson, J. Purcell, R. Tousey, Bull. Amer. Phys. Soc. 29, No. 4 (1954).

5. D. Stasey, G. Stith, R. Nidey, W. Pietenpol, Electronics 27, 149 (1954); R. Nidey, D. Stasey, Rev. Sci. Instr. 27, 216 (1956).

6. W. Pietenpol, W. Rense, F. Walz, Phys. Rev. 90, 156 (1953).

7. W. Rense, Phys. Rev. 91, 299 (1953).

8. F. S. Jonson, J. Purcell, R. Tousey, Bull. Amer. Phys. Soc. 29, No. 4, 33 (1954); Astronom. J. 60, 165 (1955).

9. S. Miller, R. Mercure, W. Rense, Astrophys. J. 124, 580 (1956).

10. R. Mercure, S. Miller, W. Rense, F. Stuart, J. Geophys. Res. 61, 571 (1956).

11. E. Bayram, T. Chubb, H. Friedman and N. Gailar, Phys. Rev. 91, 1278 (1953).

12. E. Bayram, T. Chubb, H. Friedman and J. Kupperron, JOSA 46, 384 (1956).

13. I. S. Shklovsky, Izvestiya Kr. AO 4, 80 (1949); The Solar Corona, Gostekhizdat, 1951.

14. R. Athay and R. Thomas, Astrophys. J. 124, 586 (1956).

15. E. Bayram, T. A. Chubb and H. Friedman, J. Geophys. Res. 61, 251 (1956).

16. G. Elwert, Zeits. f. Naturforsch. 9a, 637 (1954).

17. Science 124, 673 (1956).

18. J. Kaplan and H. Odishaw, Science 122, 1003 (1955).

19. R. Tousey, JOSA 47, 261 (1957).

20. E. T. Byram, T. A. Chubb, H. Friedman, J. Kupperian, Astrophys. J. 124, 480 (1956).

21. E. R. Mustel, A. B. Severny, Izvestiya Kr. A. O. 8, 19 (1956).