&£tHOe!HZ3
SECRETARIA DE PLANEJAMENTO DA PRESIDÊNCIA DA REPUBLICA
CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO
,11ft INSTITUTO IE PESQUISAS ESPACIAIS 1. Publication N9 2. Version 3. Date 5. Distribution INPE-2672-SRE/281 NOV. 1982 D Internal GO External 4. Origin Program D Restricted CAS RADIO 6. Key words - selected by the author(s)
SOLAR BURSTS; M4-WAVE AND HARD X-RAYS; SUB-SECOND STRUCTURES.
7. U.D.C.: S23.74S
INPE-2672-JPRE-/281 8. Title VI 10. N9 of pages: is
SUB-SECOND PULSATIONS SIMULTANEOUSLY OBSERVED11 . Last page: 24 AT MTCROWAVES AND HARD X-RAYS IN A SOLAR BURST 12. Revised by
9. Authorship / T.Takdkura* i— (T.Takakura) P.Kaufmann J. E.R.Costa 13. Authorized by S.S.Degaonkar* N.Ohki** N.Nitta oo- Nelsonton tie Jesus Paraà Responsible author ff >^%M $**SL— Director 14. Abstract/Notes - Sub-second time structures have been fount in the emissions during solar bursts in rmrwaves and,independently,in hard X-rays. However, simultaneous observations of such fast time structures in mm radio and X-ray ranges fias not been available so far. Accordingly, wt^pijamed coordinated^ observations of solar bursts in November 1981 with a high time fesolutiõrTóf a few lint/UsecoruW. The hard X-rays (30-40 keV were observed with hard X-ray monitor (HXM) aboard the Hinotori Satellite with a time resolution of 7.81 ms and the radio emissions ware observed on the ground with 4Sft dish at Itapetinga Radio Observatory with a high time resolution (1 ms) and high sensitivities at 22 GHz and 44 GHz, supplemented by a patrol observation at 7 GHz with time resolution of 100 ms. The pulsations repeated with a period of about 300 ms. The physical implication of the good correlation is not clear at this stage, but it may give a clue to the understanding of the high energy phenomena occurring during the solar flares
* Dept. of Astronomy, Univ. of Tokyo, Japan. ** Tokyo Astronomical Observatory, Univ. of Tokyo, Japan.
15. Remarks - Accepted by Nature. Supported partially by FINE?. Sub-second Pulsations Simultaneously Observed at Microwaves and Hard X-rays in a Solar Burst
T. Takakura*, P. Kaufmann**, J E.R. Costa**, S.S. Degaonkar*+, K. Ohki****, and N. Nitta* * Department of Astronomy, University of Tokyo, Bunkyo-ku, Tokyo 113, Japan ** INPE: Instituto de Pesquisas Espaciais, CP. 515, 12200, São José dos Campos, SP, Brasil *** Tokyo Astronomical Observatory, university of Tokyo, Mitaka, Tokyo 181, Japan + On. leave of absence from physical Res. Lab., Ahmedabad, India
Sub-second time structures have been found in the emis- sions during the solar bursts in mm waves 1 '2 and, independ- ently, in hard X-rays 3 '4 . However, simultaneous observation of such fast time structures in mm radio and X-ray ranges has not been available so far. Accordingly, we planned coordinated observations of solar burst in November 1981 with a high time resolution of a few milli-seconds. The hard X-rays (30 - 40 keV) were observed with hard X-rays monitor, HXM , aboard the Hinotori Satellite with a time resolution of 7.81 ms and the radio emissions were observed on the ground with 45 ft dish at Itapetinga Radio Observatory with a high time resolution (1 ms) and high sensitivities at 22 GHz and 44 GHz , supple- -2-
mented by patrol observations at 7 GHz with a time resolution of 100 ms. Absolute timing at Itapetinga and Hinotori is bet ter than 10 ms. Correlated sub-second time structures at hard X-rays and at mm-waves were found for the first time.
A solar sub-flare occurred on 1981 November 4 at IS 28*20s UT. The associated radio and X-ray bursts are shown in com pressed time scale in Figure 1 (a) and (b). As can be seen the correlation of the major burst time structures is poor among the three microwave frequencies, and between microwaves and hard X-rays. However, during this solar burst, it was found in time expanded sections that quasi-periodic pulsation.7 were present, repeating at a rate of 3-4 s""1. The pulsations per sisted throughout the maximum phase of the burst, i.e. from about 1828:10 - 1828:40 UT, at 22 GHz, 44 GHz and at hard X- rays. For the analysis with greater time resolution we expanded various sections in that time interval. As examples we show here two sections labeled (A) and (B) in Figure 1, each of 2 seconds duration. They are shown in Figure 2 (A) and (B). In Figure 2, in curves .(b) and (d) for both sections A and B, gradual component has not been subtrated. In the plots (c) and (e), the radio flux was filtered by taking the second time derivative in order to eliminate the contamination due to changes in the gradual background level. The X-ray data were not fil tered in this manner since the background level was nearly constant in those periods. The statistical error a of the -3-
X-ray counts,given by the square root of an average counting rate, shown in Figure 2(a) is large and corresponds to about 50 percent of the peak-to-peak amplitude of the pulsations. * The error bars of the radio signal are due to system temper ature and are small as indicated by o in Figure 2(b) and (c).
We have labeled by numbers 1 to 6 the most distinctive fast time structures found in Figure 2 for sections A and B. It is clear that an inspection of the time structures indi cates a striking correspondence amongst each other. For most of the time structures, there is nearly one-to-one correspond ence. The peak-to-peak amplitude of the X-ray pulsations is about 30 percent of the average counts, and is larger than the statistical error. On the other hand, the peak-to-peak ampli tude of the radio pulsations is about 1 percent of the gradual background level, and is larger than the system noise error bars. The best signal-to-noise is obtained at 22 GHz.
In order to test the significance of the correlation, cross-correlation analysis was made. The cross-correlation coefficients are given by
1 -n-J+1 c (x x) (v (1] j - H^r J1 i' i+j-i" y> '
. where x, y are the mean values for the time interval under investigation. The errors for the coefficients are derived -4-
from the general theory of propagation of errors ,
cj k axk xk ayk yk which reduces from equation (1) to
The cross-correlations between X-rays and radio fluxes are shown in Figure 3 for the sections A and B (cf. Figs. 1, 2). The abscissa indicates delay time (in sec) of radio emis sion from X-rays. In Figure 3(a) and (c), the cross-correla tions for the raw radio data without filtering are shown and in Figure 3(b) and (d), those with filtered radio data are shown. It can be seen that the influence of the gradual back ground at 22 GHz and 44 GHz is significant. In Figure 4 (a) and (b), we have shown the cross-correlation between filtered 22 GHz and 44 GHz data for the sections A and B. The accuracy in time delays for these plots of cross-correlations is 62.5 ms.
The quasi-periodicity is evident as the repetition of cross-correlation coefficient peaks occurs in all the plots of Figures 3 and 4. The cross-correlation coefficient plot for 22 GHz and 44 GHz data (Figure 4) is the best for quasi- periodicity determinations. It gives a period of about 290 -5-
ms for section A, and of about 320 ms for section B. Periods of the order of 300 ms are also evident on the plots of cross- correlation coefficients in Figure 3.
The 22 GHs; and 44 GHz pulsations appear to be essentially in phase (to within 62.5 ms) for both sections A and B (Fig. 4). The hard X-rays pulsations, however, do not seem to be time coincident with the mm-wave pulsations. From Figure 3(A), it appears that 22 GHz peaks might be delayed with respect to the hard X-ray peaks, by about 200 ms and the 44 GHz peaks ap pear to be delayed by about 250 ms with respect to hard X-ray peaks, or in phase, or advanced by a small amount (within the time inaccuracy of 62.5 ms). Similarly, in Figure 3(B), the mm-wave peaks, both at 22 and 44 GHz, might be delayed by 250 ms with respect to hard X-rays, or in phase, or advanced by a small amount (within the time inaccuracy of 62.5 ms). It is more likely that the nun-wave peaks are delayed by an amount close to one period (i.e., about 300 ms).
Thus, the foregoing observations have revealed an inter esting correlation of quasi-periodic sub-second pulsations in ram-waves and hard X-rays. The physical implication of the good correlation with a suggested phase shift of the order of one-period is not clear at the present stage, but it is de finitely linked with the understanding of the high energy phenomena occurring during the solar flares. 6-
If we suppose that the pulsations are attributed to the MHD waves moving with a speed of the order of 10' •5 kms-1, the 300 ms period gives the size of about 103 km for the source showing modulation in intensity. On the other hand, the X-ray image of the flaring site obtained with the hard X-ray tele- scope SXT .' aboard the Hino tori indicate a single source of about 20" arc (about 1.5*10* km) in HWHM during the section A of Figure 1. These measurements have spatial and temporal resolutions of about 10" arc and 7 sec, respectively. The, source may, however, be composed of many unresolved small sources as it has been proposed by some models ' . A number of situations might be conceived. If only one of the smaller sources is suffering a modulation in intensity, the number of elements instantaneously present must be less than three in order to satisfy such a large modulation as 30 percent. On the other hand, quasi-periodic reconnections due to tearing 13 mode instability in a magnetic tube has been suggested by Spicer The multiple sources, however, need not necessarily be inter dependent or sequential. They could be attributed to an ag gregation of many disrupting flux tubes superimposed on a grad- 12 ual longer disruption, as has been recently suggested . The superposition would sometimes result in the apparent quasi- periodic fluctuations, whose amplitude depends on the pulse waveform in response to each individual disruption and on how they pile up. Another alternative could be conceived by attributing the -7-
period of 300 ms to the bouncing motion of a cloud of high energy electrons in a magnetic flux tube of the order of 10 km in length. A recent model assuming MHD oscillations at the flaring site produces pulsations of microwave and X-ray 14 emission . In this model fluxes of energetic electrons are modulated at the top of a loop, producing hard X-rays at every bounce by bremsstrahlung lower in the chromosphere and modula tion of microwaves is attributed to the growth rate of Bernstein mode. For producing modulation at such high frequencies as 22 GHz and 44 GHz, however, this model reguires very high plasma density of the order of IO12 ^ 1013 cm""3 at the top of the coronal magnetic tube.
It is worthwhile to add that fast microwave time structures superimposed on the event of 4 November 1981 as reported here were also observed at 10.6 GHz at Algongwin Radio Observatory, Canada . This data will be included in a more extended study of this event to be published in the future.
The authors wish to acknowledge the contribution from the late Dr. P.M. Strauss (1942-1981) to the developments at Itapetinga which made the present high sensitivity observations possible. Deepest thanks of the authors T.T., K.Ô., and N.N. are due to the staff of the Institute of Space and Astronaut- . ical Science, especially to the members of Hinotori team for devoting to every stage, from the design to the launching and -8-
operation of the Hinotori Satellite. Another author (S.S.D.) thanks Prof. T. Takakura and JSPS for their kind hospitality. Acknowledgements are due to F.N.S. Carvalho for help in Itapetinga data processing.
References
Kaufmann, P.; Strauss, F.M.; Opher, R. and Laporte, C.,
Astron. Astrophys., 87f 58 (1980). 2 Hurford, G.J.; Marsh, K.A.; Zirin, H.; Kaufmann, P. and Strauss, F.M., in preparation (1982). Abstract in Bull. Am. Astron. Soc, 11, 678 (1979). Orwig, L.E.; Frost, K.J. and Dennis, B.R., Astrophys. J., 244, L163 (1981). 4 Charikov, Y.E.; Kocharov, G.E. and Lazutkov, V.P., 17th Int. Cosmic Ray Conf., Paris, 13-25 July (1981). 5Ohki, K.; Nitta, N,; Tsuneta, S.; Takakura, T.; Makishima, K.; Murakami, T.; Ogawara, Y. and Oda, M., Proceeding of Hinotori Symposium on Flares (published by ISAS, 1982). Kaufmann, P.; Strauss, F.M.; Schaal, R.E. and Laporte, C, Solar Phys., 78, 389 (1982). 7Bevington, P.R., "Data Reduction and Error Analysis for the Physical Sciences", McGraw-Hill, New York, USA (1969). -9-
Makishima, K., Proceeding of Hinotori Symposium on Flares (published by ISAS, 1982). Takakura, T.; Tsuneta, S.; Ohki, K.; Nitta, N.; Makishiroa, K.; Murakami, T.; Ogawara, T.; Oda, M. and Miyamoto, S., Astrophys. J., submitted (1982). Takakura, T.; Ohki, K.; Tsuneta, S. and Nitta, N., Proceeding of Japan-US Seminar on Recent Advances in the Understanding of Solar Flares (1982). Brown, J.C.; Craig, J.D. and Karpen, J.T., Solar Phys., 67, 143 (1980). * 12 Sturrock, P.A.; Kaufmann, P.. and Smith, D.F., "Energy Release in Solar Flares", Stanford Univ. Inst. Plasma Res., Kept. SUIPR No.933, Stanford, USA, October (1982). 14 Zaitsev, V.V. and Stepanov, A.V., Solar Phys., submitted (1982) Tapping, K.F., private communication (1982).
Captions to the figures
Fig. 1 - Time profiles in microwaves (a) and in hard X-rays (b) for the solar burst of 4 November 1981. Vertical lines show intervals labeled (A) and (B) which are expanded in Figs. 2 and 3, showing the superimposed 3-4 s"1 quasi-pulsations. Fig. 2 - Expanted time profiles for 2 sec in intervals (A) and (B) with an accuracy of 62.5 ms. Displayed plots -1 fl
are for 1.875 sec. Fig. 3 - Normalized cross-correlation coefficients between X-rays and microwaves in intervals (A) and (D) for time delays extending to ±1 sec with an accuracy of 62.5 ms. Fig. 4 - Normalized cross-correlation coefficients between 22 GKz and 44 GHz in intervals (A) and (B) for time delays extending to ±i sec with an accuracy of 62.5 ms. » t/MOV. 1181_ XTfijPETiHG-Pt
: I I V 7£H?. I V"7
LVtfyiAAAAJUipOAMJlJUUU^^ 1622:2ò\. — -.3? I
/* 2^.'20 nzvoo
X-**» •"•*- *+">-**»£ ccoij-cc>KX£i^Tè(uJ CosrríCêC*i3 J£en'ot* 4-
fr H-O
X-A*r < **&(* AVIIMO JX-RW » 31 fHt FitJXAKO
(h ~At>
M-/.0. -A2
*-*&* *V£tfj. X-*AX wW&Ht- i-O
*I.O
1-1.2 •• . I -IO I X-fiVff- ttCHt PiLTEuàl> l-O '• f-P-
? &>
•ti'O e)
C»~) w
F''&> 3 22GtJi VS.
•S£er*'o*s /)
C») -to
x-/., \
'SEcrJD/sJQ
Lb
â&úiy&Ecj
P7 G-.tf '