3 Results

   
3.1 Radio continuum images

 

 

Table 3: Basic properties and imaging parameters.

Galaxy	$\alpha$	$\delta$	D	i a	${f_{60}}\over{f_{100}}$	Beamb	$1\,\sigma$ rms	S1.43	Pos. Angle
	(2000)	(2000)	(Mpc)	($^\circ$)		FWHM ('')	(mJy beam-1)	(mJy)	PA ($^\circ$)
NGC1055	02 41 43.7	+00 25 54	16.0	67	0.34	54	0.08	$233\pm10$	$102.6\pm0.2$
NGC1406	03 39 22.5	-31 19 19	14.9	90	0.44	43	0.08	$127\pm\,~7$	$16.2\pm0.3$
NGC1421	03 42 29.5	-13 29 23	31.1	90	0.40	$57\times37.5(30)$	0.10	$110\pm\,~7$	$0.5\pm0.1$
NGC1511	03 59 35.7	-67 38 07	17.5	72	0.62	17.7c/33.0d	0.070c/0.05d	$156\pm\,~8$	$122.8\pm0.2$
NGC2748	09 13 44.6	+76 28 32	28.7	71	0.39	52	0.10	$66.4\pm6$	$42.1\pm0.5$
NGC2820	09 21 44.1	+64 15 26	29.9	90	0.33e	$62\times41(75)$	0.06	$59.0\pm8$	$67.4\pm0.2$
NGC3175	10 14 42.3	-28 52 20	15.9	85	0.46	$43\times33(52)$	0.07	$69.9\pm6$	$51.2\pm0.2$
NGC3437	10 52 34.9	+22 56 04	25.5	75	0.57	$70\times41(-71)$	0.10	$73.6\pm6$	$111.6\pm0.1$
NGC3717	11 31 31.8	-30 18 32	27.1	90	0.44	41	0.08	$123\pm10$	$33.7\pm0.6$
NGC4527	12 34 08.5	+02 39 11	31.9	68	0.47	$55\times50(67)$	0.15	$235\pm10$	$68.3\pm0.2$
NGC4700	12 49 07.3	-11 24 46	25.5	90	0.57	$67\times46(50)$	0.10	$29.6\pm3$	$49.8\pm0.8$
NGC5073	13 19 20.8	-14 50 35	43.9	87	0.69	$71\times44(0)$	0.15	$38.9\pm4$	$176.4^f\pm0.4$
NGC7090	21 36 28.6	-54 33 26	11.7	90	0.33	17.7c/34.5d	0.075c/0.05d	$38.0\pm4$	$128.2\pm0.2$
NGC7462	23 02 46.5	-40 50 07	15.1	90	0.49	23.5c/46.5d	0.085c/0.05d	$26.0\pm3$	$69.7\pm1.0$
NGC7541	23 14 43.0	+04 32 05	42.4	75	0.48	47	0.15	$158\pm\,~8$	$95.6\pm0.4$

Notes to Table 3:
^a) Inclination angle; approximate values based on optical axial ratios, from Tully (1988). In particular the value for NGC1055 appears to be too low, possibly because of its prominent bulge.
^b) Major axis $\times$ minor axis (position angle), if not circular.
^c) At 2.45 GHz observing frequency.
^d) At 1.43 GHz observing frequency.
^e) The FIR flux measurements of NGC2820 are contaminated by contributions from its partners, NGC2820A and NGC2814.
^f) NGC5073 is basically unresolved; the radio PA reflects the beam shape. During the data analysis PA $= 56^\circ$ was used.

The 12 galaxies observed with the VLA are displayed in Figs. 1-3, 5-12 and 15. For these objects only 1.425 GHz data are currently available.

ATCA radio continuum images of the three galaxies NGC1511, NGC7090 and NGC7462 at 2.45 GHz and 1.43 GHz are presented in Figs. 4, 13 and 14. Each figure displays the 2.45 GHz image on the left side and the 1.43 GHz map on the right.

In the following we briefly describe the radio images of each galaxy in our small sample of 15 systems. Note that the first positive contour always displays the $2\sigma$ confidence level and further contours increase by factors of $\sqrt {2}$ from there.

The angular resolution and sensitivities (1$\sigma$) of the maps are listed in Table 3. We also include in Table 3 the measured total 1.43 GHz flux densities, S_1.43, and the position angles of the radio continuum emission distribution, as determined from fitting two-dimensional ellipsoidal Gaussians to the radio continuum emission distributions, using the AIPS task JMFIT.

NGC1055:   The 1.425 GHz radio continuum emission of NGC1055 is displayed in Fig. 1. It is centrally peaked, with resolved disk emission out to radii of about $4\hbox{$.\mkern-4mu^\prime$ }5$. As in the subsequent figures, the full width at half maximum (FWHM) of the beam is displayed in the lower left corner. There are several nearby point sources, which are most likely unrelated.

NGC1406:   Like NGC1055, NGC1406 (Fig. 2) exhibits centrally peaked emission from its disk, however with little substructure at the angular resolution of the current data.

NGC1421:   NGC1421 is resolved by our observations, which were combined with data by Irwin et al. (1999). The emission is centrally peaked, but the central maximum is not very bright (Fig. 3). Several nearby unrelated point sources had to be removed for investigations of its structure perpendicular to the disk plane.

NGC1511:   NGC1511 (Fig. 4) has an unusual radio continuum distribution in its disk. There are two prominent maxima, about 30'' apart. The surface brightness of the emission is high at both 2.45 GHz and 1.43 GHz. Perpendicular to the disk plane the emission distribution is very thick. The total height of the emission distribution at 2.45 GHz is about 5 beamwidths (about 4 at 1.43 GHz). The axial ratio in the optical is 2.9 (NASA Extragalactic Database; NED), while the radio axial ratio is approximately 1.4 (at 2.45 GHz) and even 1.2 at 1.43 GHz.

None of the radio continuum emission from NGC1511 is associated with SN1935C (or "Nova'' Hyi 1935; van den Bergh & Hazen 1988), which occurred further to the north-east. Also the two sources about 2' south of NGC1511 are probably unrelated.

NGC2748:   The radio continuum emission distribution of NGC2748, displayed in Fig. 5, is centrally peaked but shows almost no significant substructure at the current resolution. The source to the west of NGC2748 is unrelated.

NGC2820:   Figure 6 shows a complex emission distribution, comprised of contributions from NGC2820 (east of the centre) and NGC2814 (to the west) as the major peaks and NGC2820A as a tertiary maximum in between these two. Fitting Gaussians to the emission peaks of these sources, a fourth component remains that is associated with the companion of NGC2814. Another member of this galaxy group, NGC2805 (outside the displayed field of view) was also detected.

This proximity of the different emission sources makes it difficult to disentangle the contributions from each of the systems. Using Gaussian ellipsoids to fit the individual emission distributions, we succeeded in separating the radio emitters. However, the same problem of flux contamination also applies to the FIR data, based on which the target was selected (see Table 4), where the sources cannot be spatially separated due to the lower angular resolution of IRAS compared with our data.

NGC3175:   The radial extent of NGC3175's radio continuum emission distribution (Fig. 7) is small compared to its optical extent (Condon et al. 1996). The emission is centrally peaked and marginally resolved along the disk plane, with no directly visible substructure. The source to the north-west is unrelated.

NGC3437:   NGC3437 (Fig. 8) shows little substructure in its radio continuum emission distribution at the present resolution.

NGC3717:   The radio continuum emission distribution of NGC3717 is complex (Fig. 9). Besides a strong central peak extended disk emission is visible as an emission ridge along the major axis, with a secondary emission peak about 3' south-west of the centre and another, weaker source about 3' north-east of the nucleus. The total radial extent of the emission distribution is about 7'. South-east of the centre, further emission is visible away from the disk plane, which is not easily removable by subtracting a point source profile and might therefore be genuinely extended. The weak point source north-west of the centre is probably unrelated.

Figure 2: VLA 1.43 GHz radio continuum contour map of NGC1406, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.23, -0.16, 0.16 (=$2\sigma$), 0.23, 0.32, 0.45, ..., 40.1 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

Figure 3: VLA C+D array 1.465 GHz radio continuum contour map of NGC1421, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.4, -0.28, -0.2, 0.2 (=$2\sigma$), 0.28, 0.44, 0.56, 0.8, ..., 35.2 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

NGC4527:   The radio continuum emission of NGC4527, as displayed in Fig. 10, is bright, centrally peaked and extended along its major axis, with an extent of about $6\hbox{$.\mkern-4mu^\prime$ }5$. The ratio of major-to-minor axis extent is less than 2, which is low for a highly inclined galaxy. The fact that the minor axis extent of the emission distribution is about 4 beamwidths indicates that it is resolved in this direction, despite the fact that there is little visible substructure. The central source, which is classified as an H II/LINER nucleus (NED), was subtracted for further processing (see Sect. 3.2).

NGC4700:   Despite its small angular extent, the radio continuum emission distribution of NGC4700 (Fig. 11) is resolved in both the major and minor axis direction. The axial ratio of the emission distribution is only 1.25, which is extremely low. This is a clear indication of extraplanar emission, especially when the elliptical shape of the beam, with its major axis aligned with that of the galaxy, is taken into account.

Two strong point sources east of NGC4700 (one just outside the displayed field of view) are unrelated. A third, weaker point source about $2\hbox{$.\mkern-4mu^\prime$ }5$ south-west of the galaxy centre might possibly be related with NGC4700, being located near its disk plane, however beyond the radial limit of the optical emission distribution. This point source was not included in the total flux measurement and when determining the position angle, PA (Table 3).

NGC5073:   The emission of NGC5073 is unresolved (Fig. 12), with no immediately visible sign of extraplanar emission at the angular resolution of the current data.

NGC7090:   NGC7090 (Fig. 13) exhibits centrally peaked radio continuum emission from its disk. The radial extent of the disk emission is about 4' at 2.45 GHz and $5\hbox{$.\mkern-4mu^\prime$ }5$ at 1.43 GHz. The disk emission is clearly resolved, with secondary maxima most clearly visible in the 2.45 GHz map.

There are also clear indications of extraplanar emission from cosmic ray electrons (CRs) in the halo. The halo emission is detected and spatially resolved at both observing frequencies, 2.45 GHz and 1.43 GHz. The 1.43 GHz halo emission is apparently more extended and clearly brighter than at 2.45 GHz.

Figure 4: ATCA radio continuum maps of NGC1511. Left panel: 2.45 GHz map; right panel: 1.43 GHz map; both are overlaid on a DSS optical image. The contours display the -4, -2.8, -2, 2, 2.8, 4, ... $\sigma$ confidence level (the 1-$\sigma$ rms is 0.07/0.05 mJy beam-1 at 2.45/1.43 GHz, respectively).

Figure 5: VLA 1.43 GHz radio continuum contour map of NGC2748, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.4, -0.28, -0.2, 0.2 (=$2\sigma$), 0.28, 0.44, 0.56, 0.8, ..., 35.2 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

Figure 6: VLA 1.43 GHz radio continuum contour map of NGC2820, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.17, -0.12, 0.17, 0.24, ..., 30.7 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

 

 

Table 4: Measurements of the exponential disk and halo z scale heights.

Galaxy	Obs.	Radiala	Resolution	Disk		Halo
	Freq.	Range	$\perp$ Major Axis	$z_{\rm0,disk}$	Rel.	$z_{\rm0,halo}$(''/kpc)			Rel.
	(GHz)	(''/kpc)	(''/kpc)	(''/kpc)	Ampl.	Side1b	Side2b	Average	Ampl.
NGC1055	1.425	$\pm150$/$\pm11.6$	54.0/4.19	19.2/1.49	1.00	--	--	--	0.00
NGC1406	1.425	$\pm\,~66$/$\pm\,~4.8$	43.0/5.28	4.8/0.35	0.85	27.3/1.97	18.6/1.34	23.0/1.66	0.15
NGC1421	1.465	$\pm\,~90$/$\pm13.6$	37.5/5.66	13.3/2.01	1.00	--	--	--	0.00
NGC1511	1.43	$\pm\,~34$/$\pm\,~2.9$	33.0/2.80	10.5/0.89	0.96	14.9/1.27	26.6/2.26	20.7/1.76	0.04
	2.45	$\pm\,~60$/$\pm\,~5.1$	17.7/1.50	7.9/0.67	0.90	17.9/1.52	18.1/1.54	18.0/1.53	0.10
NGC2748	1.425	$\pm\,~78$/$\pm10.9$	52.0/7.24	12.5/1.74	1.00	--	--	--	0.00
NGC2820	1.425	$\pm\,~42$/$\pm\,~6.1$	41.0/5.95	10.8/1.57	1.00	--	--	--	0.00
NGC3175	1.425	$\pm\,~54$/$\pm\,~4.2$	33.0/2.55	10.3/0.80	0.94	20.3/1.57	0.0/ 0.0	10.2/0.79	0.06
	1.49c	$\pm\,~46$/$\pm\,~3.5$	15.0/1.16	5.3/0.41	0.84	27.6/2.13	14.4/1.11	21.0/1.62	0.16
NGC3437	1.425	$\pm\,~66$/$\pm\,~8.2$	41.0/5.07	9.8/1.21	1.00	--	--	--	0.00
NGC3717	1.425	$\pm\,~90$/$\pm11.8$	41.0/5.39	14.2/1.87	1.00	--	--	--	0.00
NGC4527	1.425	$\pm114$/$\pm17.6$	50.0/7.74	20.8/3.22	1.00	--	--	--	0.00
NGC4700	1.425	$\pm\,~78$/$\pm\,~9.7$	46.0/5.69	0.1/0.01	0.33	26.9/3.33	22.5/2.78	24.7/3.06	0.67
NGC5073	1.425	$\pm\,~66$/$\pm14.1$	44.0/9.37	14.7/3.13	1.00	--	--	--	0.00
NGC7090	1.43	$\pm\,~74$/$\pm\,~4.2$	34.5/1.96	7.0/0.40	0.58	24.5/1.39	38.0/2.16	31.2/1.77	0.42
	2.45	$\pm124$/$\pm\,~7.0$	17.7/1.00	2.5/0.14	0.59	21.0/1.19	29.0/1.65	25.0/1.42	0.41
NGC7462	1.43	$\pm\,~54$/$\pm\,~4.0$	46.5/3.41	10.0/0.73	0.65	25.3/1.85	24.4/1.79	24.8/1.82	0.35
	2.45	$\pm\,~76$/$\pm\,~5.6$	23.5/1.72	2.5/0.18	0.66	26.2/1.92	20.6/1.51	23.4/1.71	0.34
NGC7541	1.425	$\pm\,~54$/$\pm11.1$	47.0/9.67	9.8/2.02	1.00	--	--	--	0.00

Notes to Table 4:
^a) Range (in units of arcsec and kpc) parallel to the galaxy's major axis over which data were averaged.
^b) "Side 1'' is always the left-hand side in the plots in Figs. 18-33, "Side 2'' the right-hand side.
^c) Data from Condon et al. (1996).

Figure 7: VLA 1.43 GHz radio continuum contour map of NGC3175, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.2, -0.14, 0.14 (=$2\sigma$), 0.2, 0.28, 0.4, ..., 24.64 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

Figure 8: VLA 1.43 GHz radio continuum contour map of NGC3437, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.28, -0.2, 0.2 (=$2\sigma$), 0.28, 0.44, 0.56, 0.8, ..., 35.2 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

NGC7462:   Figure 14 displays centrally peaked emission in the disk of NGC7462 at both 2.45 GHz and 1.43 GHz. The 2.45 GHz map exhibits a secondary emission maximum about $1\hbox{$.\mkern-4mu^\prime$ }3$ east of the centre and a weaker corresponding maximum to the west of the nucleus.

The 2.45 GHz map also shows some emission south of the centre, outside the disk of NGC7462. At 1.43 GHz one can discern extended emission from the halo. Despite the different angular resolutions of the two maps it appears that the halo is more prominent at 1.43 GHz than at higher frequencies. The 2.45 GHz map may also lack sensitivity and thus not show the full extent of the emission at this frequency yet. Note that the total flux densities of this galaxy at 2.45 GHz and 1.43 GHz are quite low (Table 3).

NGC7541:   NGC7541 (Fig. 15) was observed in bad weather conditions, which caused phase problems that could not be removed entirely from the visibility data using the self-calibration technique. The 1.43 GHz emission distribution is centrally peaked and marginally resolved along the major axis, without visible substructure at the current resolution. The source to the south-west of NGC7541 is NGC7537. The other nearby point sources are probably also unrelated.

   
3.2 Investigations of the z structure

In order to investigate the z structure of the sample galaxies (i.e., the emission distribution perpendicular to the disk plane), the following technique was applied (as done by us earlier; see e.g. Hummel et al. 1991b; Dahlem et al. 1994; Dahlem et al. 1997):

Figure 9: VLA 1.43 GHz radio continuum contour map of NGC3717, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.16, 0.16 (=$2\sigma$), 0.23, 0.32, 0.45, ..., 56.7 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

Figure 10: VLA 1.43 GHz radio continuum contour map of NGC4527, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.42, -0.3, 0.3 (=$2\sigma$), 0.42, 0.6, 0.84, ..., 105.6 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

1.

First, nearby confusing point sources were removed, using the AIPS routine JMFIT for two-dimensional ellipsoidal Gaussian fitting of the source profiles;

2.

then the unresolved nuclear source was fitted and removed from those galaxies where the nuclear emission contributes $\geq$25% of the total flux density. This applies to NGC1055, NGC1406, NGC3717 and NGC4527;

3.

Subsequently the radial range was defined over which radio data were averaged and then an average cut through the galaxy, parallel to its minor axis, was produced over that radial range (called a "z profile'');

4.

The peak surface brightness in this z profile was normalised to unity and the resulting graph displayed as a solid line in the plots introduced below;

5.

A cut through a Gaussian representing the shape of the restored beam along the galaxy's minor axis was created and its peak also normalised to unity. This Gaussian (the beam shape) represents the instrumental response to an infinitesimally thin disk at perfect edge-on inclination of $90^{\circ}$. It is shown in all plots of z profiles as a bold line;

6.

An exponential function was created, modelling the z structure of a (thin) disk, which was also normalised in its maximum to unity. The width of this function may represent the intrinsic thickness of the disk, widening of the emission distribution of an infinitely thin disk due to non-perfect edge-on viewing geometry, or a superposition of both;

7.

This exponential was then convolved with the beam and the width of the resulting distribution adjusted to mimic the width of the central maximum in the observed z profile;

8.

In case this results in a good fit to the data, no further processing is required and it is assumed that no halo emission was detected;

9.

In those cases where residual emission, in particular "wings'' in the z profiles, cannot be fitted by a single beam-convolved exponential function, a second, wider beam-convolved exponential was added to the model. This second component mimics the behaviour of that part of the emission distribution that we associate with a radio halo;

10.

Due to asymmetries in the observed emission distributions, the exponential scale heights of this wide emission component were allowed to be different on either side of the disk plane, with a common maximum in the disk plane to ensure continuity of the model distribution;

11.

This two-component model was then fitted to the data. The relative amplitudes of both components are free fit parameters, with the constraint that the sum of both amplitudes be 1;

12.

The following profiles were then plotted together for each galaxy, on a logarithmic scale:

(a)

The observed z profile (solid line),

(b)

the restored beam (bold line),

(c)

the beam-smeared exponential thin disk model (dotted line),

(d)

the beam-smeared exponential halo model (second dotted line; only if required by fit), and

(e)

the sum of both components (i.e. the "best fit'' model to the data; dashed line).

The relevant parameters and results of this procedure are collected in Table 4. The graphs described above for each of the 15 galaxies are reproduced in Figs. 18-33. The 1.43 GHz data of NGC7090 serve as a showcase to demonstrate the above steps.

Figure 11: VLA 1.43 GHz radio continuum contour map of NGC4700, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.28, -0.2, 0.2 (=$2\sigma$), 0.28, 0.44, 0.56, 0.8, ..., 35.2 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

NGC7090:   The relatively high resolution of our data of NGC7090 allows for a clear distinction between disk and halo emission in this galaxy. Therefore, we use it here as a showcase to demonstrate the process of modeling the z distribution of the emission. The observed emission distribution is reproduced on a linear scale as the solid line in the left panel of Fig. 16. An exponential disk model was created (narrow peaked distribution, solid line) and convolved with the beam (bold line), leading to the smoother profile (dotted line). Its amplitude is initially fixed at 1 to see whether it can fit the observed emission distribution. However, deviations from the disk model set in at about 25-30% of the peak surface brightness. In such cases, when residuals remain, a second, wider exponential is added and beam-convolved. The two components are then summed and the resulting function (dashed line) fitted to the data. For easier inspection of the exponential wings, the data and model, including its individual components, are then reproduced on a logarithmic scale (right panel of Fig. 16). One can see that the halo emission is asymmetric, with brighter emission towards positive offsets (the north-east) than south-west of the disk. To save space, only logarithmic plots are shown for all other galaxies. Deviations from a purely exponential decay are visible at 2.45 GHz (Fig. 17). These might be caused by the spurs with relatively high surface brightness detected by Harnett & Reynolds (1985). Such details are visible, because the 2.45 GHz map of NGC7090 is not only one of two maps with the highest angular resolution in the sample, but-because of the proximity of NGC7090-also the spatial resolution is good (1 kpc). The peak surface brightness of the averaged z profile at 2.45 GHz is only 4 mJy beam^-1 and this map is accordingly the one with the smallest dynamic range (Fig. 17). Thus, the data used for calculating $z_{\rm0,halo}$ go down only to the 5%-level (log = -1.3) and the influence of noise is visible in the wings of the emission distribution.

Figure 12: VLA 1.43 GHz radio continuum contour map of NGC5073, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.6, -0.42, -0.3, 0.3 (=$2\sigma$), 0.42, 0.6, 0.84, ..., 26.4 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

Figure 13: ATCA radio continuum maps of NGC7090. Left panel: 2.45 GHz map; right panel: 1.43 GHz map; both are overlaid on a DSS optical image. The contours display the -4, -2.8, -2, 2, 2.8, 4, ... $\sigma$ confidence level (the 1-$\sigma$ rms is 0.075/0.05 mJy beam-1 at 2.45/1.43 GHz, respectively).

Figure 14: ATCA radio continuum maps of NGC7462. Left panel: 2.45 GHz map; right panel: 1.43 GHz map; both are overlaid on a DSS optical image. The contours display the -4, -2.8, -2, 2, 2.8, 4, ... $\sigma$ confidence level (the 1-$\sigma$ rms is 0.085/0.05 mJy beam-1 at 2.45/1.43 GHz, respectively).

Figure 15: VLA 1.43 GHz radio continuum contour map of NGC7541, superimposed on a Digital Sky Survey (DSS) image. The contour levels displayed are -0.6, -0.42, -0.3, 0.3 (=$2\sigma$), 0.42, 0.6, 0.84, ..., 76.8 mJy beam-1, with a contour spacing of a factor of $\sqrt {2}$.

   
3.2.1 The z structure of the disk emission

The values of the "disk scale height'', $z_{\rm0,disk}$, in Table 4 represent the apparent exponential scale height of the thin disk emission component. At the resolution of our data, and given the uncertainties in the values of the inclination angles, i, it is not clear what causes the apparent thickness of the thin disk component in each galaxy. The total width could be due to either an inclination of $i < 90^\circ$ or intrinsic disk thickness, or a superposition of both. The  $z_{\rm0,disk}$ values are thus not in all cases a direct measure of the intrinsic disk thickness.

A comparison of the best-fitting model of a thin disk with the observed cut perpendicular to the disk plane tells immediately whether there is any residual emission that cannot be explained by the single-component model or not. Any emission at high projected z offsets that is not well represented by the best-fitting single exponential (dotted line) must arise from outside the galaxy disk. In the presence of significant amounts of residual emission, especially in the form of broad wings in the emission distribution perpendicular to the disk plane, it is assumed that a radio halo exists. For all practical purposes, it is assumed that the galaxy disks are not warped in those parts from where the observed radio continuum emission arises. In a number of cases there is only marginal or no evidence for the existence of extraplanar synchrotron radio continuum emission. Some of our maps do not have sufficient resolution to draw firm conclusions yet. This affects in particular the most distant objects in the sample (NGC5073 and NGC7541), but also a few others.

   
3.2.2 Search for radio halos and measurement of their properties

In those cases where there is excess emission beyond the disk, the scale height of the second, wider exponential component is interpreted as the exponential scale height of the halo emission. It has been noted by different authors before that the surface brightness of radio halos declines exponentially away from the galaxy disks (see, for example, the high-resolution z profile through NGC891 by Dahlem et al. 1994). Therefore, the halo thickness can be quantified quite naturally by the exponential scale height of the halo emission, $z_{\rm0,halo}$, as listed in Cols. 7-9 of Table 4. All other relevant parameters, including the relative amplitudes of the exponential components fitted and the radial ranges (Col. 3) over which the fits to the disk and halo emission components were performed, are tabulated there as well. Note that the relative amplitude of the two exponential components is no indication of the actual total flux percentages, because the z profiles were performed only over a limited radial range, excluding the outer parts of the galaxy disks.

Data for the fits from which the exponential z scale heights were determined (as listed in Table 4) were used only above the noise level of the images (Figs. 18-33). The threshold in all cases lies in the range from 1-5% of the peak surface brightness, depending on data quality and source brightness. In the following we will present a short description of the observed z structure of all galaxies in the sample (except NGC7090, which was already introduced above).

Figure 16: Cut through the 1.43 GHz emission of NGC7090, perpendicular to the disk plane, on a linear scale (left) and logarithmic (right). In the linear graph the observed data (solid line), beam profile (bold line), exponential disk model and exponential halo model (solid lines, peaked distributions), the two beam-convolved exponential components (dotted lines) and the total model (sum of the two exponential components; dashed line) are presented. The logarithmic plot shows the same components, except the initial unconvolved exponentials.

NGC1055:   There is no evidence in Fig. 18 for the presence of emission beyond the thin disk of NGC1055. The beam-convolved exponential disk model is a good representation of the data. Therefore, the dotted line of the single-exponential component model and the dashed one of the total model overlap exactly, resulting in a dash-dotted line that in turn conincides in large parts with the solid line representing the data.

NGC1406:   Although relatively weak, the emission beyond the best-fitting disk model in the logarithmic plot in Fig. 19 is statistically significant and can be fitted only by adding a second exponential component. Higher angular resolution is required to confirm this preliminary result and to allow for a more reliable calculation of $z_{\rm0,halo}$.

NGC1421:   The z profile of NGC1421 (Fig. 20) does not give any indication of the presence of wings that cannot be fitted by a single-exponential model.

NGC1511:   Figures 21 and 22 display the z profiles at 2.45 GHz and 1.43 GHz, respectively. Both provide evidence for the existence of extraplanar radio continuum emission. The slopes of the halo emission are about linear in the logarithmic plots, indicating the expected exponential decay. Although the relative amplitude of the halo emission at 1.43 GHz is only 4% of the peak, it is clearly detected (Table 4). The significantly lower amplitude of the halo component at 1.43 GHz compared to 2.45 GHz is most likely caused by beam smearing of disk emission.

NGC2748:   Figure 23 illustrates how precisely the emission profile of NGC2748 can be approximated by a single beam-smeared exponential distribution.

NGC2820:   No proof of excess emission beyond the best-fitting disk model is visible in our data of NGC2820 at the present resolution (Fig. 24).

NGC3175:   Although not very extended and at a low level, deviations from the best-fitting disk model were found on the south-west side of NGC3175, which are modeled by an additional exponential component displayed in Fig. 25. In Sect. 4.2.1 radio data with a higher angular resolution are presented that clearly confirm the presence of a radio halo.

Figure 17: Cut through the 2.45 GHz emission of NGC7090, perpendicular to the disk plane (same line types as the logarithmic plot in Fig. 16).

Figure 18: Cut perpendicular to the disk plane of NGC1055, perpendicular to the disk plane (same line types as in Fig. 17).

NGC3437:   There is no evidence for the existence of extraplanar emission from NGC3437 from the z profile in Fig. 26. Some low surface brightness excess emission is visible on the north side (positive offsets). However, the result is not clear enough to be considered a secure detection because of residual phase uncertainties in the data. Observations with higher resolution and better signal-to-noise ratio are warranted.

NGC3717:   NGC3717 exhibits no deviations from the best-fitting disk model in Fig. 27, but note the large scale height of the emission distribution compared to those galaxies where disk and halo emission could be separated.

NGC4527:   In the case of NGC4527 insufficient angular resolution is not the primary problem in the detection of halo emission (see Fig. 28). However, it has the widest apparent disk emission distribution of all objects in the sample, probably due to its relatively low inclination (see Col. 5 in Table 3). No significant emission beyond the disk model is detected. The slope of the measured data is partly even steeper than that of the model (especially on the north side), indicating that there is either a residual data problem or that the emission distribution falls off more steeply with increasing z-distance than an exponential.

NGC4700:   The most convincing new detection of a radio halo amongst the galaxies observed with the VLA D array has been made in NGC4700, a starburst galaxy that had scarcely been studied prior to our observations. The logarithmic z profile in Fig. 29 exhibits wide linear wings on both sides of the galaxy disk, indicating the presence of an exponentially decaying CR halo that can be traced reliably out to the 1.25% level, up to $1\hbox{$.\mkern-4mu^\prime$ }5$ (10 kpc) away from the disk plane, with an unusually high scale height of 3.1 kpc. Deviations from the best fit to the disk emission start at a very high surface brightness. In fact, NGC4700 is the only galaxy in the present sample in which the halo exponential has a higher amplitude than the disk component. The scale height of the disk component is not well-constrained because it is unresolved.

NGC5073:   This is the most distant object in our sample. No halo emission could be detected from this galaxy and the z profile of the disk is also almost unresolved by the beam (Fig. 30). The fact that the observed profile crosses the beam profile might indicate residual phase problems in the data or a slope that is steeper than exponential.

Figure 19: Cut perpendicular to the disk plane of NGC1406 (same line types as in Fig. 17).

Figure 20: Cut perpendicular to the disk plane of NGC1421 (same line types as in Fig. 17).

Figure 21: Cut through the 2.45 GHz emission of NGC1511, perpendicular to the disk plane (same line types as in Fig. 17).

Figure 22: Cut through the 1.43 GHz emission of NGC1511, perpendicular to the disk plane (same line types as in Fig. 17).

Figure 23: Cut perpendicular to the disk plane of NGC2748 (same line types as in Fig. 17).

Figure 24: Cut perpendicular to the disk plane of NGC2820 (same line types as in Fig. 17).

Figure 25: Cut perpendicular to the disk plane of NGC3175 (same line types as in Fig. 17).

Figure 26: Cut perpendicular to the disk plane of NGC3437 (same line types as in Fig. 17).

Figure 27: Cut perpendicular to the disk plane of NGC3717 (same line types as in Fig. 17).

Figure 28: Cut perpendicular to the disk plane of NGC4527 (same line types as in Fig. 17).

Figure 29: Cut perpendicular to the disk plane of NGC4700 (same line types as in Fig. 17).

Figure 30: Cut perpendicular to the disk plane of NGC5073 (same line types as in Fig. 17).

Figure 31: Cut through the 2.45 GHz emission of NGC7462, perpendicular to the disk plane (same line types as in Fig. 17).

Figure 32: Cut through the 1.43 GHz emission of NGC7462, perpendicular to the disk plane (same line types as in Fig. 17).

Figure 33: Cut perpendicular to the disk plane of NGC7541 (same line types as in Fig. 17).

NGC7462:   Clear signatures of extraplanar emission are detected in the z profiles through NGC7462 at both observing frequencies, 2.45 GHz (Fig. 31) and 1.43 GHz (Fig. 32). Also at both frequencies an asymmetry of the halo is apparent, the brighter emission being observed on the south side (towards negative offsets).

NGC7541:   This galaxy, at almost the same distance as NGC5073, is also unresolved along its minor axis by our observations, as displayed in Fig. 33. No excess emission is observed beyond the exponential disk model.

   
3.2.3 Summary of observations

1.

Halo emission was detected unambiguously in 6 galaxies out of 15. The "raw'' detection rate is therefore 40%;

2.

the values of the exponential scale heights, $z_{\rm0,halo}$, of the detected radio halos range from 1.4 to 3.1 kpc;

3.

the data of all other galaxies can be approximated by a single-component exponential z distribution model.

Our 1.43 GHz (and 2.45 GHz) radio continuum observations are therefore a suitable tool for searches of gaseous halos around spirals.

   
3.3 Radio spectral indices

 

 

Table 5: Radio spectral indices from ATCA data.

Galaxy	2.45 GHz	$\bar\alpha^{2.45}_{1.43}$
	flux (mJy)	( $S \propto \nu^{\bar\alpha}$)
NGC1511	$101\pm5$	$-0.81\pm0.18$
NGC7090	$24\pm3$	$-0.86\pm0.43$
NGC7462	$16\pm2$	$-0.91\pm0.45$

From the VLA observations, and thus for 12 of the 15 galaxies, no spectral index information is available. However, the ATCA observations at two frequencies simultaneously do provide as a by-product spectral indices, $\alpha^{2.45}_{1.43}$. We have derived mean spectral indices, $\bar\alpha^{2.45}_{1.43}$, over the entire galaxies, as tabulated in Table 5. Mean values around $\alpha\ = -0.8$, as observed in the three galaxies NGC1511, NGC7090 and NGC7462, are quite normal for late-type spiral galaxies (e.g., Lisenfeld & Völk 2000 and references therein).

Images of the spectral index distribution within the three galaxies are displayed in Figs. 35-37. All three displays are presented in the same fashion, i.e., with the same grey scales and contours (see figure captions), making them directly comparable. The angular resolution of the spectral index maps is that of the 1.43 GHz data. Spectral indices are displayed only in those areas where the signal-to-noise ratio in both input images (1.43 and 2.45 GHz) is $\geq$$5\,\sigma$. The boundaries surrounding the spectral index distributions indicate that the maps were blanked outside this area.

The spectral index distribution in NGC1511 (as displayed in Fig. 35) is quite homogeneous, with a few edge effects (especially at the southern and northern end of the distribution) that should not be taken at face value.

In NGC7090 (Fig. 36) the flattest spectrum (i.e., most visibly influenced by thermal emission) is observed in the central part of the disk, with a slight steepening towards the outer disk and the halo regime.

NGC7462 (Fig. 37) exhibits a spectral index steepening slightly away from the disk plane, with little substructure at the resolution of the current data.

A general trend is visible that the spectral indices steepen slightly away from the disk planes, indicating that energy losses of the CRs play an important role.

   
3.4 Magnetic fields

The maps in Figs. 1-15 show only total emission because, with the current sensitivity and angular resolution, no polarised radiation was detected (ATCA data) or calibration not performed (VLA observations). Thus, no information is available on magnetic field configurations in the observed galaxies. This implies that any possible effect that the magnetic field configuration might have on the observed CR distribution in the halos cannot be investigated here.

Figure 34: Cut perpendicular to the disk plane of NGC3175 (same line types as in Fig. 17). Based on VLA C array data by Condon et al. (1996).

Figure 35: 2.45 GHz vs. 1.43 GHz spectral index distribution within NGC1511. The grey-scale ranges from -2 to 1; the contours start at -2, increasing by 0.2 each. The central oval contour represents a spectral index of -0.8.

Figure 36: 2.45 GHz vs. 1.43 GHz spectral index distribution within NGC7090. The grey-scale ranges from -2 to 1; the contours start at -2, increasing by 0.2 each. The central closed contour represents a spectral index of -0.6.

Figure 37: 2.45 GHz vs. 1.43 GHz spectral index distribution within NGC7462. The grey-scale ranges from -2 to 1; the contours start at -2, increasing by 0.2 each. The central contours closest to the disk mid-plane represent a spectral index of -0.6.