Dust coma of comet C/2009 P1 (Garradd) by imaging polarimetry



Comet C/2009 P1 (Garradd) was observed by imaging polarimetry for nearly 5 months from October 2011 to March 2012, over an intermediate phase angle range (28°–35°). Two months before perihelion and one month after, dust particles seem to be ejected all around the optocenter and jets extend to distances greater than 40,000 km. An increase of activity is noticed in intensity and polarization after perihelion. Two months before perihelion and one month after, the dust emission seems to be all around the optocenter. Two and three months after perihelion the jets are mainly toward the solar direction with an extension of more than 20,000 km projected on the sky. The values of the aperture polarization are comparable to those of other comets. On the polarization maps in October 2011 and January 2012 the higher polarization zones extend in large regions perpendicularly to the solar direction where jets are also observed. In February and March 2012, the polarization in the jets is larger in the solar direction than in the surrounding coma. By its activity visible on intensity images and polarization maps at large distances from the nucleus, comet Garradd probably belongs to the high-Pmax class of comets.


Comets have progressively been recognized as able to provide key information about the origin and early evolution of our solar system. Some of them are likely to have formed in the vicinity of giant planets before being stored much farther away in the Oort cloud and eventually returning as new active comets. Other ones are likely formed in the Kuiper Belt and transported in the inner solar system as periodic comets (e.g., Festou et al. 2004). Amongst new bright comets, comet C/2009 P1 Garradd (hereafter Garradd) was discovered in August 2009 by G. J. Garradd from Siding Springs Observatory in Australia at 8.7 AU from the Sun. This seemingly dynamically new comet presents an almost parabolic orbit. It reached its perihelion on 23 December 2011, at a solar distance (R) of about 1.5505 AU. Earth-based observations were favorable until early March 2012, where its distance to the Earth (Δ) went down to about 1.27 AU. Ground-based observations (Villanueva et al. 2012) and space-based observations from, e.g., Swift satellite and Deep Impact space probe (Bodevits et al. 2012) have indicated a large dust production rate, even far away from the Sun, suggesting that its activity might not only be triggered by water ice sublimation.

Cometary dust, ejected by sublimating ices gas drag into the low-density gaseous comae is efficiently studied by observations of solar scattered light and specifically from its partial linear polarization (e.g., Dollfus et al. 1988; Levasseur-Regourd et al. 1999; Hadamcik and Levasseur-Regourd 2003a). Such measurements are monitored as a function of (α), the phase angle between the directions of the Sun and of the observer as seen from the scattering dust and (λ), the wavelength of the observations, allowing comparisons of the dust properties between different regions of a given coma and between different comets. Comparisons to experimental and numerical simulations provide clues to the dust properties (Hadamcik et al. 2006, 2007a Levasseur-Regourd et al. 2007; Lasue et al. 2009; Das and Sen 2011). We have observed comet Garradd in the context of a France-India research project. This collaboration, allowing polarimetric observations of the same small object from two different observing sites in the Northern hemisphere, Haute-Provence Observatory (OHP) in France and Girawali Observatory in India, has already provided results about, e.g., comet 67P/Churyumov-Gerasimenko, to better characterize its environment in preparation of the Rosetta mission that is to rendezvous with its nucleus in 2014 (Hadamcik et al. 2010). Comet Garradd observations, from October 2011 to March 2012, corresponded to increasing solar distances (R) from 1.33 to 1.93 AU, to decreasing Earth distances (Δ) from 2.10 to 1.33 AU, and to phase angles remaining between 28° and 35°, i.e., in a region where differences between cometary dust presenting a low polarization and cometary dust presenting a high polarization are already detectable.

We first describe the instrumental techniques, including the data reduction, and present the log of the observations. Then, we analyze the results about the brightness of the light scattered by dust and its decrease as a function of optocentric distance. The linear polarization is analyzed in terms of aperture polarization (allowing easy comparisons with other comets) and of polarization maps, before discussing the significance of these results for the properties of comet Garradd'dust.

Observations and Data Reduction

Two telescopes with a Cassegrain configuration were used: a 80 cm telescope at OHP in France and a 2 m telescope at IUCAA Girawali Observatory (IGO) in India. At OHP, four polaroids are mounted on a rotating wheel; the polarized components with their fast axis at 45° from one another are recorded successively. At IGO, a rotating half-wave plate and a Wollaston prism allow us to record two perpendicularly polarized components on the same CCD image (frame): in the next image the two components are polarized at 45° from the previous ones (for more details see Sen and Tandon 1994; Ramaprakash et al. 1998). Different continuum filters avoid or reduce contaminations by the gaseous emissions (Table 1). For more details on the instruments and data reduction, see, e.g., Hadamcik et al. (2010, 2013).

Table 1. Instrumentation at OHP and IUCAA
ObservatoryTelescope diameter apertureCCD resolution (binning)PolarizerCometary continuum filters





0.8 m


0.21 arcsec/pixel

4 × 4 pixels

Rotating 4 polaroids

ESA narrow band

CB 443 nm, Δλ 4 nm

CR 684 nm, Δλ 9 nm


ROHP 650 nm, Δλ 90 nm

IOHP 810 nm, Δλ 150 nm





2 m


0.307 arcsec/pixel

1 × 1 pixel

Rotating half-wave plate + Wollaston prism

ESA narrow band

CB 443 nm, Δλ 4 nm

CR 684 nm, Δλ 9 nm


RIGO 630 nm, Δλ 120 nm

IIGO 900 nm, Δλ 80 nm

Four polarized components are obtained for each series of observations. The intensity (I), polarization (P), and polarization angle (θ) are calculated by the following expressions:

display math(1)
display math(2)
display math(3)
display math(4)

The polarized intensities are measured in the instrumental reference system. θr =  − θ0) is the position angle of the polarization plane in the equatorial reference system. The position angle of the scattering plane (ϕ) is known at each date. For symmetry reason it is defined between 0° and 180° and can be deduced from the value of the Sun-comet radius vector position angle (Sun-C PA in Table 2). The sign between the parentheses, in Equation (4), is chosen to ensure the condition 0° < (ϕ ± 90°) < 180°. θr is generally of about 90° for comets observed at phase angles larger than 25° (Levasseur-Regourd et al. 1996).

Table 2. Log of the observations
DateΔ (AU)R (AU)α (°)Sun-C PA (°)FiltersResolution (km/pixel)
  1. Δ = distance to Earth; R = solar distance; α = phase angle; Sun-C PA = extended Sun-comet radius vector position angle (JPL Horizons ephemeris).

IGO, October 21–22, 20112.10–2.111.6130.9–30.869–68CB, CR470
OHP, October 26, 20112.111.6030.364ROHP1290
OHP, January 23–25, 20121.68–1.641.60–1.6134.8–35.2317–315CR, ROHP1030–1000
IGO, February 18–20, 20121.35–1.341.7434.7–34.3285–281CR, RIGO, IIGO300
OHP, March 17–19, 20121.37–1.331.96–1.9328.5–28.3172–165CB, CR, ROHP, IOHP840–810

A center gravity algorithm is used to find the position of the optocenter of the comet on each polarized image. To avoid artifacts in the polarization maps, the polarized components are centered with a precision of better than 0.1 pixel before any calculation. The sky background is estimated in a region outside the coma and subtracted. Then, fluxes through apertures with diameters of 12 and 24 pixels centered on the optocenter are measured for each polarized components (I0, I45, I90, and I135), and their stability in each aperture is controlled as well as the total intensity I (Equation (1)) for the whole series. If a difference greater than 2% is detected between the fluxes measured on the successive images, the image is rejected.

The log of the observations is presented in Table 2, together with the filters used for each run. During the October 2011 observations the sky was partially cloudy at both sites, for OHP observations on a run of 4 nights, only 4 h were acceptable. In January and February 2012, the comet was only observable during 2 or 3 h at the end of the nights. During each night, polarized and unpolarized stars were observed to estimate the residual instrumental polarization and to determine the origin of the instrumental reference system θ0 (Table 3).

Table 3. Polarized and unpolarized standard stars
 StarP (%)PA (°)Pobs (%)PAobs (°)θ0
  1. Observations in red (R), blue (B), and (V) filters. P and PA from the literature (Turnshek et al. 1990; Schmidt et al. 1992), Pobs and PAobs from observations, θ0 = (PA − PAobs). NA = not available.

OHP OctoberHD236633 (R)5.376 ± 0.02893.04 ± 0.155.25 ± 0.1092 ± 11
HD21447 (B)0.017 ± 0.0328.63 ± 0.30.02 ± 0.02(R)25 ± 44
BD+59°389 (R)6.43 ± 0.0298.1 ± 0.026.20 ± 0.16103 ± 2−5
HD251204 (R)4.27 ± 0.028147 ± 24.32 ± 0.10138 ± 109
OHP JanuaryHD21447 (B)0.017 ± 0.0328.630.02 ± 0.0524 ± 55
G191B2B0.09 ± 0.061470.02 ± 0.02129 ± 1018
HD236633 (R)5.376 ± 0.02893.04 ± 0.215.41 ± 0.0598 ± 8−5
HD94851 (B)0.057 ± 0.018NA0.163 ± 0.067.5 ± 5'
HD94851 (V) NA0.07 ± 0.0589 ± 5 
IGO FebruaryHD25443 (R)4.734 ± 0.045133.65 ± 0.284.854 ± 0.1132.28 ± 11.5
HD25443 (V)5.127 ± 0.061134.23 ± 0.345.1 ± 0.2132.5 ± 52.3
BD+59°389 (R)6.43 ± 0.02298.14 ± 0.106.4 ± 0.1105 ± 8−7
BD+59°389 (V)6.701 ± 0.01598.09 ± 0.076.5 ± 0.191 ± 87
HD94851 (B)0.06 ± 0.02NA0.08 ± 0.1091 ± 8 
OHP MarchHD155197 (R)4.274 ± 0.027102.88 ± 0.184.1 ± 0.2114 ± 51
GD319 (B)0.045 ± 0.047140.790.07 ± 0.1161 ± 10−21
GD319 (V)0.089 ± 0.093140.150.08 ± 0.1 (R)135 ± 10 (R)5
HD1551974.38 ± 0.031034.05 ± 0.1593 ± 1010


Intensity variations and coma morphologies are presented from October 2011 to March 2012. Then, variations of the linear polarization are discussed.

Intensity Images

The polarized components are added to build the intensity images for each filter (Equation (1)). The azimuthally averaged radial log-log profile is studied first (Fig. 1). For all the runs, the slope of decrease is nominal (−1 ± 0.05) from the optocenter as expected from an isotropic coma (out of the blurred central region by seeing). This nominal (−1) slope at large optocentric distances indicates that the sky background has been correctly subtracted for each series of observations.

Figure 1.

Azimuthally averaged radial intensity profiles for the four periods of observations and fits with a (−1) slope.

The overall shape of the coma is circular in projection on the sky with a limited sunward-tailward asymmetry. Figure 2 presents isophotes levels in the left column. Intensity images analyzed by a rotational gradient method (Larson and Sekanina 1984)—to emphasize the azimuthal intensity gradients generally corresponding to jet structures—are presented in the right column. For all observing periods, the main feature is more developed in the projected solar direction. In projection, it extends over more than 40,000 km. During October, the jets present several fan structures. One fan structure with PA ≈ [240°–310°] is close to the solar direction (PA = 244°) while there is another one in the antisolar direction with PA ≈ [50°–130°]. In January, the jets seem to be curved anticlockwise all around the optocenter with a slightly larger extension in the solar direction (PA = 135°). Close to the optocenter three different ejecting directions are noticed with PA ≈ 110°, PA ≈ 5°, and PA ≈ 325° (not easily seen on the gray-scale image). During February, structures are well observed in the solar direction with fan-shaped fine jets between PA ≈ [85°–120°] (solar PA = 105°) and on the two sides shorter jets at PA ≈ 55° and PA ≈ 115°. Projected on the sky, the main jets extend over more than 22,000 km long. The structures are similar during the three nights of observations through the cometary red and infrared filters. During March observations, a V-shaped structure is observed. Its longer branch is oriented in the solar direction, with an extension of more than 40,000 km at a PA ≈ 340°. The shortest branch has an extension of more than 12,000 km on the sky with a PA ≈ 35°. Close to the optocenter the ejection seems to be mainly in the South direction (about antisolar). For the three nights of observation in March, similar structures are observed with all the filters.

Figure 2.

Intensity images for October, January, February, and March. Left column: isophotes with the solar direction indicated by a line with image. Right column: images of intensity treated by a rotational gradient method (in negative). Fields 45,000 × 45,000 km. North is up and East is left.

Figure 3a compares the radial decrease profiles in intensity between the solar and antisolar directions as a function of the optocentric distance for the four periods. The profiles present similar overall shapes with a higher intensity in the solar direction up to more than 25,000 km from the optocenter. The slope values can be found in Table 4. The slope increases progressively with the optocentric distance in the solar direction, indicating that dust is moving from this region. A slope smaller than −1 indicates that dust is pushed tailward by the solar radiation replenishing the antisolar regions. Figure 3b and Table 4 compare the decrease profiles in intensity between the two perpendicular directions to the solar direction. The differences between these profiles are very small, as generally observed.

Table 4. Slopes in log–log scale in the solar (S), antisolar (anti-S) directions, and in the perpendicular (⊥) directions for the four runs of observationsThumbnail image of
Figure 3.

a) Profiles as a function of optocenter distance for comparison of intensity decrease between the solar and antisolar directions for each period of observations. b) Profiles as a function of optocenter distance for comparison of intensity decrease between the perpendicular directions to the solar directions for each period of observations.

The changing solar distance can explain the change in the structures as compared to the solar directions between the four runs, with more activity after perihelion (January and February). It also depends on different geometric configurations related to the orientation and projection of the jets as seen from the Earth.

Linear Polarization

Aperture polarization is obtained for each period and filter (Equation (2)). The values are compared with values on the synthetic phase curves for other comets. On the polarization maps, regions with a higher polarization than the surrounding coma are detected, suggesting differences in the physical properties of the dust.

Aperture Polarization

The fluxes are measured through different apertures on the polarized component images (sum of all the selected images for each orientation of the polarization). The polarization is calculated using Equation (2). It also provides the determination of the integrated polarization whenever the cometary tracking is imperfect. The results are presented in Table 5 and in Fig. 4A. The position angle of the polarization plane in the equatorial reference system (related to the solar PA) θr is on average about 90° (calculated by Equations (3) and (4) in an aperture of about 50 pixels diameter) for all the observations, without any variation close to the optocenter. The wavelength dependence is presented in Fig. 4B. As expected for such phase angles no significant differences are observed between the polarization values outside the error bars.

Table 5. Polarization values (in %) for different apertures and wavelengths. Error bars correspond to a confidence level of 95% (2σ)
Date 2011–2012Day (UT)Aperture diameter filter5000 km10,000 km20,000 km30,000 km40,000 kmθr (°)
IGO October (α = 30.9°–30.8°)21CB2.6 ± 0.42.5 ± 0.52.8 ± 0.72.3 ± 0.8NA88 ± 10
22CR2.9 ± 0.32.8 ± 0.22.8 ± 0.62.6 ± 0.6NA 
 CB2.8 ± 0.52.7 ± 0.42.3 ± 0.62.1 ± 0.8NA 
 CR2.7 ± 0.52.6 ± 0.42.5 ± 0.52.5 ± 0.6NA 
OHP October (α = 30.3°)26ROHP2.6 ± 0.52.8 ± 0.42.8 ± 0.42.7 ± 0.42.6 ± 0.494 ± 10
OHP January (α = 34.8°–35.2°)23ROHPNA3.3 ± 0.54.1 ± 0.44.5 ± 0.54.2 ± 0.689 ± 10
24CRNA4.7 ± 0.24.7 ± 0.34.2 ± 0.64.1 ± 0.675 ± 10
25CR5.7 ± 0.44.8 ± 0.34.7 ± 0.34.7 ± 0.33.8 ± 0.585 ± 8
IGO February (α = 34.7°–34.3°)18RIGO3.6 ± 0.33.7 ± 0.33.7 ± 0.33.7 ± 0.43.6 ± 0.588 ± 5
 CR3.7 ± 0.23.6 ± 0.33.6 ± 0.33.5 ± 0.53.3 ± 0.599 ± 10
19IIGO3.9 ± 0.23.8 ± 0.23.6 ± 0.33.5 ± 0.43.2 ± 0.597 ± 10
20IIGO3.8 ± 0.33.9 ± 0.23.8 ± 0.23.7 ± 0.23.6 ± 0.588 ± 10
 CR3.9 ± 0.33.7 ± 0.43.7 ± 0.43.6 ± 0.53.5 ± 0.5 
 RIGO3.9 ± 0.23.9 ± 0.33.8 ± 0.33.8 ± 0.43.5 ± 0.5 
 CR3.9 ± 0.33.7 ± 0.43.6 ± 0.53.7 ± 0.53.2 ± 0.6 
OHP March (α = 28.5°–28.3°)18ROHP2.2 ± 0.32.1 ± 0.22.2 ± 0.22.1 ± 0.22.0 ± 0.389 ± 5
19CR2.1 ± 0.32.3 ± 0.32.2 ± 0.22.2 ± 0.32.1 ± 0.390 ± 3
 CR2.3 ± 0.32.25 ± 0.32.3 ± 0.22.2 ± 0.22.2 ± 0.3 
201OHP5.0 ± 0.53.2 ± 0.22.9 ± 0.22.8 ± 0.32.5 ± 0.390 ± 4
 CBNA1.9 ± 0.42.0 ± 0.41.9 ± 0.5NA 
 CBNA2.1 ± 0.42.2 ± 0.42.0 ± 0.5NA 
 CR2.1 ± 0.32.3 ± 0.32.3 ± 0.32.2 ± 0.32.1 ± 0.3 
 ROHP2.3 ± 0.32.3 ± 0.22.2 ± 0.22.1 ± 0.32.1 ± 0.3 
 IOHP4.1 ± 0.42.6 ± 0.22.4 ± 0.32.0 ± 0.32.0 ± 0.3 
Figure 4.

Polarization curves. Up: in the red wavelength domain with comparison to other comets. Down: Comparison between data obtained at different wavelengths.

Polarization Maps

Polarization maps for the four periods are presented in Fig. 5. The different regions have polarization values different only by about 1%, which is at the limit of the error bars. The structures are observed over successive nights and for different combinations of the series of polarized components. In the regions where a higher polarization is observed, a slight increase in intensity is also observed on the profiles (October: Fig. 3b(E) NW; January: Fig. 3b(F) NE, optocentric distance d < 15000 km; February: Fig. 3a(C) E solar direction; March: Fig. 3a(D) solar direction and Fig. 3b(H) W for d < 8000 km).

Figure 5.

Polarization maps: A) 2011 October (ROHP filter), B) 2012 January (RC filter), C) February (RIGO filter), D) March (IOHP filter) observations. To improve the display a Gaussian filter was applied. Optocenter +, straight line: solar direction. Color or gray scales in percent. Fields 45,000 × 45,000 km. North is up and East is left. Some images have been treated by a median filter.

  1. In October (Fig. 5A), the central region presents a polarization of ≈2.6% and increases slightly in the North–West direction shaping an extending arc structure (Sun is about South-West) with a polarization of ≈3%. The difference is small but this structure is observed in different combinations of the series of images and seems to be real.
  2. In January (Fig. 5B), the higher polarization region extends North–East from PA ≈ 20° to PA ≈ 120°, close to the solar direction. The maximum polarization value is about 6% while in the South-West opposite direction it is about 4%.
  3. In February (Fig. 5C), the central region presents a polarization value between 3% and 5% at the center. It extends on average up to 7000 km distance. The polarization in the surrounding region has an apparent width around the central region of about 12,000 km with a value of approximately 3% and decreases farther away. Whenever jets appear on the treated intensity image (East direction), they are also present on the polarization map with approximately 4% polarization (the surrounding polarization being 3%). Their extension is greater than 20,000 km.
  4. In March (Fig. 5D), the central region—with an average polarization of about 4%—is elongated (East–West) with an extension of 9000 km and a width of 4000 km. Farther away a symmetric “butterfly” shape is observable on both sides extending over 14,000 km with an average polarization value of 3%. In the outer coma, the polarization value falls down to 1%. The butterfly wings extend further toward the Sun.

Amongst the four periods of observations, the smaller solar distance and the larger phase angle are in January. This can be the reason why the polarization is slightly higher during this time. The decrease in Earth-to-comet distance allows a higher resolution in the two last periods but the activity seems to decrease with the decrease in the solar distance from January to March.

Discussion and Conclusion

Linear polarization measurements of the light scattered by dust on comet Garradd have also been performed by Kiselev et al. (2012), between August 2011 and February 2012 at phase angles of about 13°, 15°, 22°, 30°, 31°, 32°, and 37°. The polarization values perfectly agree with our results and complement them.

The polarization values remain constant when the aperture is sufficiently large before decreasing at the dust coma limit (Table 5). This behavior is usually observed in comets. To point out similarities and differences in the dust ejected by comet Garradd and by other comets, the polarization values are compared with observations of other comets. From the available data in the literature, synthetic polarization phase curves were built in different wavelength domains in the visible (Levasseur-Regourd et al. 1996). Two classes were defined: (1) with a low maximum in polarization of about 10–15%, (2) with a higher polarization and a maximum of about 25–30%. Trigonometric fits define synthetic phase curves for the two classes and the wavelength domains (blue-green, red, near-infrared). The classification is generally possible once measurements are available at phase angles larger than 35°. The results for comet Garradd are similar to those obtained for other comets (Fig. 4A), perhaps closer to the high-Pmax comets class (but the difference between high and low-Pmax comets is small at phase angles smaller than 35°). Bodevits et al. (2012) suggest similarities between comet Garradd and comet Hale-Bopp. The polarization values of comet Hale-Bopp are much higher than those obtained for any other comets (Levasseur-Regourd 1999), including presently studied comet Garradd indicating at least different physical properties for their dust.

New and periodic comets have both been observed by imaging polarimetry at low or intermediate phase angles. High-Pmax comets present generally well structured jets with a higher polarization than the surrounding coma and extending at relatively large distances from their nucleus as compared with the coma size. The presence of submicron-sized grains, in highly porous large aggregates, has been suggested in the jets and arcs of comet Hale-Bopp (Hanner et al. 1997; Hayward et al. 2000; Hadamcik and Levasseur-Regourd 2003b). More generally, the particles in the jets of high-Pmax comets seem to be made of similar grains and structures but the number of the jets is relatively smaller. In the coma of low-Pmax comets, structures may be absent on the polarization maps, e.g., in comet C/1989 X1 Austin (Eaton et al. 1992) or high polarization regions with small extension around the optocenter may be found. In that case, the ejected dust probably consists of large dark particles, moving slowly and eventually breaking down. This seemed to be the case for comet 9P/Tempel 1 before the Deep Impact event and comet 67P/Churyumov-Gerasimenko before perihelion (Hadamcik et al. 2007b, 2010). The jets in comet Garradd are preferentially oriented in the solar direction. On the polarization maps, jets are mainly observable in January, February, and in March. Their extension is more than 20,000 km. Their properties are consistent with a population of submicron to micron-sized grains in aggregates. In January and February the higher polarization and extension of the jets may be the result of an increase in activity after perihelion. On March maps, the higher polarization region is relatively restricted but the butterfly-shaped structure may be the result of particles breaking up after leaving the central high polarization region. Such behavior has been observed, e.g., around the main fragment of comet C/1999 S4 (LINEAR) during its complete disintegration (Hadamcik and Levasseur-Regourd 2003c).

In conclusion comet Garradd was observed during five observational runs from October 2011 to March 2012. The phase angle varies between 28° and 35°. The linear polarization of the scattered light by the dust particles is similar to those of other comets. Jet activity is observed on each period. After perihelion, in January, February, and March, higher polarization jets are apparent in the solar direction. This activity seems to confirm a classification among the high-Pmax class of comets with ejection of small submicron to micron-sized grains possibly in aggregates.


We thank Daniel Bardin for his help in the observations at OHP. We gratefully acknowledge ESA for the narrow-band filters at OHP. OHP and IUCAA are acknowledged for allocation of observation time. The authors are part of a CEFIPRA project team (India-France CEFIPRA Research Project No. 4507-1). A. C. L. R. acknowledges support from the French Space Agency (CNES).

Editorial Handling

Dr. Donald Brownlee