1: \documentclass{eas}
2:
3: \usepackage{graphicx}
4:
5: \begin{document}
6:
7: \TitreGlobal{Mass Profiles and Shapes of Cosmological Structures}
8:
9: \title{Gravitational Lensing \& Stellar Dynamics}
10:
11: \author{Koopmans,~L.V.E.}\address{Kapteyn Astronomical Institute,
12: P.O.Box 800, 9700\,AV Groningen, The Netherlands}
13: \runningtitle{Gravitational Lensing \& Stellar Dynamics}
14:
15: %\setcounter{page}{23}
16: % Keep this line, even if the page will be settled afterwards..
17: %\index{Author1, A.}
18: %\index{Author2, B.}
19: %\index{Author3, C.}
20: % Repeat the authors here, this will help to make the final index
21:
22: \begin{abstract}
23: Strong gravitational lensing and stellar dynamics provide two
24: complementary and orthogonal constraints on the density profiles of
25: galaxies. Based on spherically symmetric, scale-free, mass models, it
26: is shown that the combination of both techniques is powerful in
27: breaking the mass-sheet and mass-anisotropy degeneracies. Second,
28: observational results are presented from the Lenses Structure \& Dynamics
29: (LSD) Survey and the Sloan Lens ACS (SLACS) Survey collaborations to
30: illustrate this new methodology in constraining the dark and stellar
31: density profiles, and mass structure, of early-type galaxies to
32: redshifts of unity.
33: \end{abstract}
34:
35: \maketitle
36:
37: \section{Introduction}
38: Strong gravitational lensing has matured over the last decades in to a
39: powerful tool in astrophysical and cosmological studies. Whereas
40: research initially focused on cosmology, in particular on the
41: determination of cosmological parameters, and conducting large surveys
42: to find new strong lens systems, emphasis is now shifting toward the
43: use of strong lenses in the study of the mass distribution of galaxies
44: and their environments (see e.g.\ Kochanek et al.\ 2004). This became
45: feasible in the 1990s because of the advent of powerful new space and
46: ground-based telescopes for imaging and spectroscopic follow-up of
47: many newly discovered lens systems.
48:
49: The combination of high-resolution multi-colour optical imaging with
50: spectroscopic redshifts, and sometimes with high-resolution
51: radio-imaging and more recently X-ray observations, of these lens
52: systems, is now providing a very detailed picture of the lens galaxies
53: and allows their mass, some-times mass profile, to be constrained.
54: However the mass-profile degeneracy introduces a limit to the accuracy
55: with which mass-profiles of strong-lens galaxies can be determined.
56:
57: On way to break this degeneracy is to have {\sl independent}
58: information on the lens mass distribution, on a scale different
59: from the Einstein radius. One particularly powerful measure of galaxy
60: mass, or its density profiles, is obtained through stellar kinematics.
61: The latter has been particularly productive in the study of the mass
62: distribution in nearby galaxies. However, the high-quality data
63: required to break for example the mass-anisotropy degeneracy has severely
64: limited similar studies of distant galaxies, even with the availability of
65: 8--10\,m class telescopes.
66:
67: The combination of strong gravitational lensing and stellar-kinematic
68: constraints -- the focus of this review -- provides a way to break the
69: above-mentioned degeneracies in both methods {\sl and} overcome the
70: additional limit set by the high S/N-ratio required for the
71: stellar-kinematic data. In Section 2, using scale-free spherical
72: models, it is shown how the combination of lensing and dynamics can
73: break existing model degeneracies. In Section 3, the results from two
74: observational programs (LSD and SLACS) are highlighted regarding
75: constraints on the dark and stellar mass structure of early-type
76: galaxies to $z\approx 1$. In Section 4, a short summary and
77: forward look is given.
78:
79:
80: \section{Spherical Scale-Free Lensing \& Stellar-Dynamical Models}
81:
82: Despite the complexity of real galaxies, studying spherically
83: symmetric scale-free galaxies turns out to be enlightening in touching
84: upon the strengths and weaknesses of combining gravitational lensing and
85: stellar-kinematic constraints.
86:
87: \subsection{Combining Lensing Mass and Stellar Velocity Dispersions}
88:
89: Let us suppose the following spherical scale-free model for the lens
90: galaxy:
91: \begin{equation}
92: \left\{
93: \begin{array}{lcl}
94: \nu_l(r) & = & \nu_{l,0} r^{-\delta} \medskip \\
95: \nu_\rho(r) & = & \nu_{\rho,0} r^{-\gamma'}\medskip \\
96: \beta(r) & = & 1-{\langle\sigma^2_\theta\rangle}/
97: {\langle \sigma^2_r} \rangle
98: \end{array}
99: \right.,
100: \end{equation}
101: where $\nu_l(r)$ is the luminosity density of stars -- a trace
102: component -- embedded in a total (i.e.\ luminous plus dark-matter)
103: mass distribution with a density $\nu_\rho(r)$. The anisotropy of the
104: stellar velocity ellipsoid is $\beta$, constant with radius. For a
105: lens galaxy with a projected mass $M_{\rm E}$ inside the Einstein
106: radius $R_{\rm E}$, the luminosity weighted average line-of-sight
107: velocity dispersion inside an aperture $R_{\rm A}$ is given, after
108: solving the spherical Jeans equations, by
109: \begin{equation}\label{eq:sigav}
110: \langle \sigma_{||}^2\rangle (\le R_{\rm A}) = \frac{1}{\pi}
111: \left[ \frac{G M_{\rm E}}{R_{\rm E}} \right] f(\gamma',
112: \delta,\beta) \times \left(\frac{R_{\rm A}}{R_{\rm
113: E}}\right)^{2-\gamma'},
114: \end{equation}
115: where
116: \begin{eqnarray}
117: f(\gamma', \delta,\beta) = 2 \sqrt{\pi}\,\left(\frac{\delta -3}{(\xi
118: - 3)(\xi - 2\beta)} \right)&\times&
119: \left\{\frac{\Gamma[(\xi-1)/2]}{\Gamma[\xi/2]} - \beta
120: \frac{\Gamma[(\xi+1)/2]}{\Gamma[(\xi+2)/2]} \right\} \times
121: \nonumber\\ && \left\{\frac{\Gamma[\delta/2]\Gamma[\gamma'/2]}
122: {\Gamma[(\delta-1)/2]\Gamma[(\gamma'-1)/2)]}\right\}
123: \end{eqnarray}
124: with $\xi = \gamma'+\delta -2$. Similarly,
125: \begin{equation}\label{eq:sigr}
126: \sigma_{||}^2(R) = \frac{1}{\pi} \left[ \frac{G M_{\rm
127: E}}{R_{\rm E}} \right] \left(\frac{\xi - 3}{\delta -3}
128: \right) f(\gamma', \delta,\beta) \times \left(\frac{R}{R_{\rm
129: E}}\right)^{2-\gamma'}.
130: \end{equation}
131: In the simple case of a SIS with $\gamma'=\delta=\xi=2$ and
132: $\beta=0$, we recover the well-known result
133: \begin{equation}
134: \sigma_{||}^2(R) = \frac{1}{\pi} \left[ \frac{G M_{\rm
135: E}}{R_{\rm E}} \right]~~~({\rm SIS}),
136: \end{equation}
137: for the Singular Isothermal Sphere mass model. At high redshifts
138: (e.g.\ $z\sim 1$), one can often only hope to measure a
139: luminosity-weighted stellar velocity dispersion of a (lens) galaxy
140: within a give aperture, even when observed with 8-10\,m class
141: telescopes. In combination, with the mass measured inside the Einstein
142: radius of a lens galaxy, how does this provide us with information
143: about the galaxy's density profile? When studying Equation
144: \ref{eq:sigav}, we see two ways $\langle \sigma_{||}^2\rangle$ is
145: changed: through the exponent ($2-\gamma'$) and
146: through a change in the normalisation~$f(\cdots)$.
147: \begin{itemize}
148:
149: \item If $R_{\rm A}$=$R_{\rm E}$, the normalisation only
150: depends on~$\gamma'$. For a given $M_{\rm E}$, the slope
151: $\gamma'$ then determines the gradient of the potential inside
152: that radius. Steeper density profiles lead to larger potential
153: gradients and, on-average, larger stellar velocities for a
154: given luminosity profile ($\delta$) and given anisotropy
155: $\beta$.
156:
157: \item If $R_{\rm A}$$\neq$$R_{\rm E}$, the stellar velocity
158: dispersion inside an aperture also depends on $\gamma'$ through
159: the factor $\left({R_{\rm A}}/{R_{\rm
160: E}}\right)^{2-\gamma'}$. Hence, roughly speaking, gravitational
161: lensing determines the mass inside $R_{\rm E}$ and stellar
162: kinematics determining the mass inside $R_{\rm A}$. The average
163: density profiles between $R_{\rm E}$ and $R_{\rm A}$ is then
164: quantified, with the stellar velocity anisotropy often a
165: secondary effect.
166:
167: \end{itemize}
168:
169: Both physical interpretations are at the basis of determining the average
170: density profile in the inner regions of gravitational-lens
171: galaxies. Real galaxies are obviously more complex, but this
172: toy-model illustrates the basic ideas of combining lensing and
173: dynamics and how it is used to measure the density profiles of (lens)
174: galaxies and break the mass-profile and mass-anisotropy degeneracies.
175:
176: \subsection{How reliable can the density profile be measured?}
177:
178: To answer this question, we take the derivative of
179: Equation~\ref{eq:sigav} with respect to $\sigma_{||}$ and $M_{\rm E}$,
180: assuming $\delta$ can be determined to good accuracy and $\beta$
181: is known or assumed. In that case, one finds
182: \begin{equation}
183: \left\langle \delta_{\gamma'}^2\right\rangle = \alpha_g^{-2}
184: \left\{\left\langle \delta_{M_{\rm E}}^2\right\rangle + 4
185: \left\langle \delta_{\sigma_{||}}^2\right\rangle \right\},
186: \end{equation}
187: with
188: $$\alpha_g \equiv \frac{1}{2}\left(\frac{\partial \log
189: f}{\partial \log \gamma'} - \gamma' \, \log\left[\frac{R_{\rm
190: A}}{R_{\rm E}}\right]\right),
191: $$ where $\delta_{\dots}$ indicate the fractional errors on the
192: respective quantities. Because in general $\delta_{M_{\rm E}}<
193: \delta_{\sigma_{||}}$, one finds the simple rule of thumb that the
194: error $\delta_{\gamma'} \sim \delta_{\sigma_{||}}$ for
195: close-to-isothermal mass models (typically one finds $\alpha_g
196: \sim 2$). We thus conclude that the fractional error on the determine
197: density slope $\gamma'$ is of the same order as the measurement error
198: on $\sigma_{||}$. This simple rule-of-thumb is confirmed for more
199: complex models as well (e.g.\ Koopmans et al.\ 2005) and typically
200: $\delta_{\gamma'} < 0.05$ for reasonable (spectroscopy) integration times
201: of $\sim$4\,hrs on 8--10\,m class telescopes for $L_*$ lens galaxies out to
202: $z\approx 1$.
203:
204: \section{Observational Programs}
205:
206: Two observational programs to combine gravitational lensing and
207: stellar kinematic constraints are described, that are currently
208: underway: (i) The Lenses Structure and Dynamics (LSD) Survey and (ii)
209: the Sloan Lens ACS (SLACS) Survey. I will mainly focus on their
210: results regarding the determination of galaxy density profiles and
211: their mass structure.
212:
213: \subsection{The Lenses Structure \& Dynamics Survey -- LSD}
214:
215: The Lenses Structure and Dynamics (LSD) Survey was started to
216: combine stellar-kinematic constraints with those from gravitational
217: lensing (e.g.\ Koopmans \& Treu 2003; Treu \& Koopmans 2004), to
218: extend E/S0 studies {\sl beyond} the local Universe.
219:
220: Analysis of the five pressure supported E/S0 galaxies in our sample at
221: $z \approx 0.5-1.0$ was presented in Treu \& Koopmans (2004), with the
222: following results: (1) Extended dark-matter halos is detected in all
223: five E/S0 galaxies at $>99\%$~CL and dark matter contributes 40--70\%
224: (15--65\%) of the total mass within the Einstein (effective) radius.
225: (2) The inner power-law slope of the {\sl total} mass distribution is
226: $\langle \gamma'\rangle =1.75\pm0.10$ (with $\rho\propto
227: r^{-\gamma'}$), close to isothermal. Including two more similarly
228: analysed systems (PG1115+080 and B1608+656) gives $\langle
229: \gamma'\rangle =1.9\pm0.1$ with 0.3 rms scatter, indicating that dark
230: and luminous mass already `conspire' at $z\approx 1$ to produce a
231: close-to-flat rotation curve, while preserving their spatial
232: segregation, similar to what is found in disk-dominated galaxies. (3)
233: The inner power-law slope of the {\sl dark-matter} mass distribution
234: is $\langle \gamma\rangle =1.3^{+0.2}_{-0.4}$ (68\% CL), consistent
235: with numerical simulations for isotropic models with little adiabatic
236: contraction at low redshifts.
237: %
238: Although these results show that good-quality stellar kinematics of
239: E/S0 lens galaxies provide powerful constraints on the inner structure
240: of individual E/S0 galaxies out to $z\approx 1$, the study of trends
241: or evolution is clearly limited by the small sample size.
242:
243: \begin{figure}[t]
244: \centering
245: \includegraphics[width=0.8\hsize]{fig1_LVEK.eps}
246: \caption{One of the newly discovered SLACS lens systems (Bolton
247: et al.\ 2005b)}
248: \label{figure_lens}
249: \end{figure}
250:
251: \subsection{The Sloan Lens ACS Survey -- SLACS}
252:
253: To increase the number of E/S0 galaxies suitable for lensing and
254: stellar dynamical studies, the Sloan Lens ACS (SLACS) Survey was
255: initiated (e.g.\ Bolton et al.\ 2005a\&b; Treu et al.\ 2005; Koopmans
256: et al.\ 2005). E/S0 lens candidates were initially selected from the
257: SDSS Luminous Red Galaxy (LRG) sample, based on the presence of
258: multiple emission lines at a redshift larger than that of the E/S0
259: galaxy (Bolton et al.\ 2004). The 49 candidates were ranked based on
260: their lensing probability and 39/49 were successfully targeted in a
261: cycle--13 HST snapshot survey through filters F435W and F814W. The
262: survey is continuing in cycle-14. Thus far a total of $>$25 new lens
263: systems have been discovered (e.g. Fig.1), making it the largest
264: sample of lens systems from a single well-defined survey.
265:
266: The main results, thus far, presented in Bolton et al.\ (2005a\&b),
267: Treu et al.\ (2005) and Koopmans et al.\ (2005) are: (1) The SLACS
268: E/S0 lens galaxy sample appears in all respects (e.g.\ colours, FP) to
269: be a random sub-sample of the LRG sample, only skewed toward the brighter
270: (i.e.\ more massive) and higher surface brightness systems, because of
271: the lensing and lensing-cross-section selection effects. The SLACS
272: sample should therefore represents the general population of massive E/S0
273: galaxies. (2) The ratio between the central stellar velocity
274: dispersion ($\sigma$) and the velocity dispersion that best fits the
275: lensing models ($\sigma_{\rm SIE}$) is $\langle f_{\rm SIE}\rangle
276: \equiv \langle \sigma/\sigma_{\rm SIE}\rangle=1.01$ with 0.065 rms
277: scatter. (3) The average logarithmic density slope for the {\sl total}
278: mass density of $\langle \gamma' \rangle = 2.01^{+0.02}_{-0.03}$ (68\%
279: C.L.) inside $\langle {\rm R}_{\rm Einst} \rangle = 4.2 \pm
280: 0.4$\,kpc. The inferred {\it intrinsic} rms spread in logarithmic
281: density slopes is $\sigma_{\gamma'}=0.12$. (4) The average
282: position-angle difference between the light distribution and the total
283: mass distribution is found to be $\langle \Delta \theta\rangle = 0 \pm
284: 3$~degrees (rms of 10 degrees), setting an upper limit of $\langle
285: \gamma_{\rm ext}\rangle < 0.035$ on the average external shear. The
286: total mass has an average ellipticity $\langle q_{\rm
287: SIE}\rangle$=0.78$\pm$0.03 (rms of 0.12), which correlates extremely
288: well with the stellar ellipticity, $q_*$, resulting in $\langle q_{\rm
289: SIE}/q_* \rangle = 0.99 \pm 0.03$ (rms of 0.11) for $\sigma >
290: 225$~km\,s$^{-1}$. Assuming an oblate mass distribution and random
291: orientations, the distribution of ellipticities implies $\langle q_3
292: \rangle\equiv \langle(c/a)_\rho\rangle =0.66$ with an error of
293: $\sim$0.2. (5) The average projected dark-matter mass fraction is
294: $\langle f_{\rm DM} \rangle =0.25 \pm 0.06$ (rms of 0.22) inside
295: $\langle {\rm R}_{\rm E}\rangle$, using the stellar mass-to-light
296: ratios derived from the Fundamental Plane as priors. (6) Combined
297: with results from the {\sl Lenses Structure \& Dynamics} (LSD) Survey
298: at $z > 0.3$, we find no significant evolution of the total density
299: slope inside one effective radius for galaxies with $\sigma_{\rm
300: ap}\ge 200$~km\,s$^{-1}$ (Fig.2): a linear fit gives $\alpha_{\gamma'} \equiv
301: d\langle \gamma' \rangle/dz=0.23\pm0.16$ (1\,$\sigma$) for the range
302: $z$=0.08--1.01 .
303:
304:
305: \begin{figure}[t]
306: \centering
307: \includegraphics[width=0.8\hsize]{fig2_LVEK.eps}
308: \caption{The density slope ($\gamma'$=$-d\log \rho/d\log r$) of
309: SLACS E/S0 galaxies (red squares) and LSD E/S0 galaxies (blue
310: circles, including PG1115+080 and CLASS B1608+656). The green
311: triangles indicate all SLACS and LSD systems with
312: $\sigma$$<$200\,km\,s$^{-1}$. The shaded region indicates
313: the $68\%$ CL around the ensemble average slope $\langle \gamma'
314: \rangle = 2.01$. Note the absence of significant evolution in
315: the density slope.}
316: \label{figure_lens}
317: \end{figure}
318:
319: \section{Conclusions}
320:
321: The joint gravitational-lensing and stellar-kinematic analysis of E/S0
322: galaxies is a new, powerful, method that allows their density profiles
323: and mass structure (among other things) to be studied beyond the local
324: Universe. Two observational programs, the LSD and SLACS Surveys, have
325: shown very promising first results, quantifying the structure and
326: evolution of galaxies for the first time out to $z=1$. This ultimately
327: allows us to directly compare with and test the hierarchical
328: galaxy-formation scenario in the $\Lambda$CDM paradigm, based not on light
329: but on mass.
330:
331: \begin{thebibliography}{}
332:
333: \bibitem[Bolton et al.(2004)]{2004AJ....127.1860B} Bolton, A.S., Burles,
334: S., Schlegel, D.J., Eisenstein, D.J., Brinkmann, J.\ 2004, AJ, 127,
335: 1860
336:
337: \bibitem[Bolton et al.(2005)]{2005ApJ...624L..21B} Bolton, A.S., Burles,
338: S., Koopmans, L.V.E., Treu, T., Moustakas, L.A.\ 2005a, ApJL, 624,
339: L21
340:
341: \bibitem[]{} Bolton, A.S., Burles, S., Koopmans, L.V.E., Treu,
342: T., Moustakas L.A. 2005b, ApJ, accepted
343:
344: \bibitem[]{}Kochanek, C.S., Schneider, P., Wambsganss, J., 2004, Part
345: 2 of Gravitational Lensing: Strong, Weak \& Micro, Proceedings of the
346: 33rd Saas-Fee Advanced Course, G. Meylan, P. Jetzer \& P. North,
347: eds. (Springer-Verlag: Berlin)
348:
349: \bibitem[Koopmans \& Treu(2003)]{2003ApJ...583..606K} Koopmans, L.V.E.,
350: Treu, T.\ 2003, ApJ, 583, 606
351:
352: \bibitem[]{} Koopmans, L.V.E., Treu, T., Bolton, A.S., Burles,
353: S.C., Moustakas L.A. 2005, ApJ, submitted
354:
355: \bibitem[Treu \& Koopmans(2004)]{2004ApJ...611..739T} Treu, T.,
356: Koopmans, L.V.E.\ 2004, ApJ, 611, 739
357:
358: \bibitem[]{} Treu, T., Koopmans, L.V.E., Bolton, A.S., Burles,
359: S.C., Moustakas L.A. 2005, ApJ, submitted
360:
361:
362: \end{thebibliography}
363: \end{document}
364:
365:
366: