The Near-Infrared Structure of NGC 1068 Jack F. Gallimore, Bucknell University Lynn D. Matthews, Harvard-Smithsonian CfA Abstract There have been several efforts attempting to resolve the Seyfert 2 nucleus of NGC 1068 in near-infrared bands. These studies have resulted in significant discrepancies in the inferred morphological and photometric properties of the nucleus. Towards resolving these discrepancies, we analyzed archival HST / NICMOS F110W, F160W, and F222M images of NGC 1068. Using a surface-brightness fitting technique, we decomposed the central 2.4” (170 parsecs at a distance of 15 Mpc) into a compact source with very red near-infrared colors (m110 – m160 = 2.9; m160 – m222 = 2.6) and an extended component with colors comparable to those of evolved stellar populations (m110 – m160 = 0.85; m160 – m222 = 0.61) that we interpret as a nuclear star cluster. The compact source makes up ~ 85% of the 2.22 micron emission within the central 2.4”, and the source size is smaller than 0.03” (2 parsecs; Thompson & Corbin 1999). The nuclear star cluster has a core diameter of ~ 50 parsecs, comparable to the distribution of CO bandhead absorption measured by Thatte et al. (1997). We consider models to explain the near-infrared colors and bolometric luminosities of the unresolved nuclear source and the nuclear star cluster. Keck 2.2 micron speckle image of NGC 1068 2.2 micron light curve of NGC 1068 (work by I. Glass) Keck Imaging vs. IR-Variability Weinberger et al. (1999) presented Keck speckle imaging data of NGC 1068, offering the highest resolution near-infrared imaging of this source to date. The photometry of the Keck speckle image (upper left) disagrees, however,  with the source size limits imposed by near-infrared variability (Glass 1995; figure above right). According to Weinberger et al., the K-band flux density of the central 70 LY is ~ 0.46 Jy. In contrast, the infrared light-curve constrains the size of a 0.46 Jy source to < 40 LY. Moreover, the nuclear point source can contribute no more variability than its 1995 flux density, ~ 0.23 Jy, limiting the size of the extended source to < 24 LY. Surface brightness slices through the NICMOS 2.22 micron image of NGC 1068 and models based on the Keck speckle image. Comparison of the Keck & NICMOS images The 2.2 micron resolution of NICMOS is ~ 0.2”, sufficient to resolve the luminous, elongated source reported by Weinberger et al. The solid line in the figure above traces a nuclear surface brightness profile from the HST / NICMOS image. The profile was taken through the nucleus and along the axis of the putative extended source. The broken lines represent models based on the Keck image and convolved with the NICMOS PSF. The models vary only in the relative contribution of the extended source: dotted line: extended flux = point flux (equivalent to the Keck photometry); short dashed line: extended flux = 50% point flux; long dashed line: extended flux = 25% point flux. Based on this profile analysis, the extended source can contribute no more than about 25% of the point source flux. Three-color image of NGC 1068 based on images in the 110w, 160w, and 222m filters. The NICMOS 2.22 micron image after substraction of the best fit compact source model. The NICMOS 2.22 micron image after subtraction of a model comprising a compact source and extended source. Radial surface brightness profiles and the best fit models.   Two-dimensional Modeling We analyzed the NICMOS images by fitting two-dimensional surface brightness models to the images of NGC 1068 (see figures above). The fitting procedure involves generating an analytical model image and convolving it with a Tiny Tim PSF. A non-linear fitting algorithm iteratively tunes the model parameters to achieve a minimum chi-square fit between the convolved model and observed images. The simple model comprising a single point source and a flat background made a poor fit to the data, and the residuals showed evidence for an extended component slightly larger than the PSF. We added an isothermal sphere model surface brightness distribution to account for this extended component, and the result was a significant improvement of the model fit. The results of the analysis are given in the tables below. Errors include statistical uncertainties and systematic uncertainties owing to focus variations and cold mask wiggle (Suchkov & Krist 1998). The flux densities of the extended and residual emission were measured within a 4.6” square aperture centered on the point source. NIC Filter Point Source  Error Extended Source Error Core Radius Error Residual (mJy) (mJy) (pc) (mJy) F110W 10.6 0.5 70.2 3.5 24.4 1.0 6.9 F160W 85.9 4.4 85.6 4.3 21.4 1.0 8.0 F222M 591.9 29.7 91.6 5.0 24.9 1.7 23.3 It is clear from the two-dimensional decomposition analysis that the point source strongly dominates at 2.2 micron and is the origin of the extremely red near-infrared colors. The size of the NICMOS point source must be < 2 pc (Thompson & Corbin 1999),  and therefore it can account for all of the variability observed by Glass (1995). The photometry of the Keck speckle image of Weinberger et al. (1999) is not compatible with the present NICMOS image. The Nuclear Point Source The figure at below left compares the near-infrared colors of the nuclear point source with the colors expected for emission from hot dust or a power law spectrum. Work in progress!