{"article_id": "0708.1996", "article_text": ["the interest in anchoring phenomena and phenomena in confined nematic liquid crystals has largely been driven by their potential use in liquid crystal display devices .", "the twisted nematic liquid crystal cell serves as an example .", "it consists of a nematic liquid crystal confined between two parallel walls , both providing homogeneous planar anchoring but with mutually perpendicular easy directions . in this case", "the orientation of the nematic director is tuned by the application of an external electric or magnetic field .", "a precise control of the surface alignment extending over large areas is decisive for the functioning of such devices .", "most studies have focused on nematic liquid crystals in contact with laterally uniform substrates . on the other hand substrate inhomogeneities", "arise rather naturally as a result of surface treatments such as rubbing .", "thus the nematic texture near the surface is in fact non - uniform .", "this non - uniformity , however , is smeared out beyond a decay length proportional to the periodicity of the surface pattern .", "very often the thickness of the non - uniform surface layer is considerably smaller than both the wavelength of visible light and the thickness of the nematic cell , i.e. , the distance between the two confining parallel walls", ". hence optical properties of the nematic liquid crystal confined between such substrates correspond to those resulting from effective , uniform substrates .", "more recent developments have demonstrated that surfaces patterned with a large periodicity of some micrometers are of considerable interest from a technological point of view ( see , e.g. , ref .", "@xcite and references therein ) .", "a new generation of electro - optical devices relies on nematic liquid crystals with patterned orientation of the nematic director over small areas which can be achieved by chemically patterning the confining surfaces .", "for example , to produce flat - panel displays with wide viewing angles one can use pixels that are divided into sub - pixels , where each sub - pixel is defined by a different orientation of the nematic director , which is induced by the surface structure and subsequently tuned by the electric field .", "in addition to the technological relevance , nematic liquid crystals in contact with non - uniform substrates provide the opportunity to study basic phenomena such as effective elastic forces between the substrates and phase transitions between various competing nematic textures ( see , e.g. , ref .", "@xcite and references therein ) .    whereas the influence of homogeneous confining substrates on nematic liquid crystals is now well understood ,", "the phase behavior of nematic liquid crystals in contact with chemically or geometrically patterned substrates is still debated .", "one might suppose that theoretical calculations based on continuum theories should resolve the properties of nematic liquid crystals in contact with patterned substrates @xcite .", "however , such calculations are numerically demanding because two- or three - dimensional grids have to be used because of the broken symmetry due to the surface pattern .", "moreover , it is very challenging to determine metastable states and energy barriers between them which are important for the understanding of bistable nematic devices @xcite . in the present paper", "we adopt a different strategy which takes the advantage of the finite decay length characterizing the influence of the surface pattern on the nematic liquid crystal in the direction perpendicular to the substrate .", "we determine first an anchoring energy function and an average surface director orientation of the patterned substrate and obtain an effective free energy for the nematic liquid crystal cell under consideration .", "we find remarkably good agreement between the phase diagrams of various systems calculated using this effective free energy function on the one hand and the original free energy functional on the other hand .", "the continuum theory for liquid crystals has its origin dating back to at least the work of oseen @xcite and zocher @xcite .", "this early version of the continuum theory for nematic liquid crystals played an important role for the further development of the static theory and its more direct formulation by frank @xcite .", "the frank theory is formulated in terms of the so - called nematic director @xmath0 , @xmath1 , and its possible spatial distortions .", "the nematic director describes the direction of the locally averaged molecular alignment in liquid crystals . in a nematic liquid crystal", "the centers of mass of the liquid crystal molecules do not exhibit long - ranged order .", "the molecules can translate freely while being aligned , on average , parallel to one another and to the nematic director .", "it is known that if an initially uniform nematic liquid crystal is distorted by external forces , it relaxes back to the uniform state after the disturbing influence is switched off , signaling that the uniform configuration represents a thermodynamically stable state .", "therefore it is assumed that there is a cost in free energy associated with elastic distortions of the nematic director of the form @xmath2&=&\\frac{1}{2}\\int_v", "d^3 r\\,\\left[k_{11}\\left(\\nabla\\cdot\\hat{{\\bf n}}\\right)^2\\right.\\nonumber \\\\&+&\\!\\!\\!\\left.k_{22}(\\hat{{\\bf n}}\\cdot(\\nabla \\times\\hat{{\\bf n}}))^2 + k_{33}(\\hat{{\\bf n}}\\times(\\nabla \\times \\hat{{\\bf n}}))^2\\right],\\nonumber \\\\ & & \\label{eq1}\\end{aligned}\\ ] ] where @xmath3 is the volume accessible to the nematic liquid crystal and @xmath4 , @xmath5 , and @xmath6 are elastic constants associated with splay , twist , and bend distortions , respectively .", "the elastic constants depend on temperature and are commonly of the order @xmath7 to . sometimes , for example , when the relative values of the elastic constants are unknown or when the resulting euler - lagrange equations are complicated , the one - constant approximation @xmath8 is made . in this case", "the elastic free energy functional reduces to @xmath9&=&\\frac{k}{2}\\int_v d^3 r\\,\\left(\\nabla\\hat{{\\bf n}}\\right)^2\\,.\\end{aligned}\\ ] ] in the presence of surfaces the bulk free energy @xmath10 must be supplemented by the surface free energy @xmath11 such that the total free energy is given by @xmath12 . in the corresponding equilibrium euler - lagrange equations @xmath13 , @xmath11 leads to appropriate boundary conditions .", "the description of the nematic director alignment at the surfaces forming the boundaries is called anchoring .", "in addition to the so - called free boundary condition where there is no anchoring , one considers weak and strong anchoring . if there are no anchoring conditions imposed on @xmath14 at the boundary , the bulk free energy @xmath15 is minimized using standard techniques of the calculus of variations . in the case of strong anchoring it", "is also sufficient to minimize the bulk free energy but subject to @xmath14 taking prescribed fixed values at the boundary . in the case of weak", "anchoring the total free energy @xmath16 , which includes the surface free energy @xmath11 , has to be minimized .", "the most commonly used expression for the surface free energy is of the form proposed by rapini and papoular by @xcite : @xmath17=\\frac{1}{2}\\int_{s } d^2 r\\,w({\\bf r } ) ( \\hat{{\\bf n}}\\cdot{\\nu})^2.\\ ] ] the integral runs over the boundary and @xmath18 is the corresponding anchoring strength that characterizes the surface .", "the local unit vector perpendicular to the surface is denoted as @xmath14 . for negative @xmath19", ", this contribution favors an orientation of the molecules perpendicular to the surface , while positive @xmath19 favor degenerate planar orientations at the surface .", "the absolute value of the anchoring strength is commonly of the order @xmath20 to .", "here we consider a nematic liquid crystal confined between a patterned substrate at @xmath21 and a flat substrate at @xmath22 where the @xmath23 axis is normal to the flat substrate . as figs .", "[ fig1 ] ( a ) , ( c ) , and ( d ) illustrate , the lower substrate is characterized by geometrical and/or chemical patterns of periodicity @xmath24 along the @xmath25 axis . moreover , the system is translationally invariant in the @xmath26 direction . within the one - constant approximation [ eq .  ( [ eq2 ] ) ] the total free energy functional is given by @xmath27=}\\nonumber \\\\&&\\frac{kl}{2}\\int\\limits_0^p dx\\ , \\int\\limits_{z_0(x)}^d dz\\ , [ \\nabla\\theta(x , z)]^2+f_s[\\theta(x , z_0(x))]\\ , , \\label{eq4}\\end{aligned}\\ ] ] where @xmath28 is the extension of the system in @xmath26 direction , @xmath29 is the surface profile of the patterned substrate , and @xmath30 . at the upper surface", "strong anchoring @xmath31 is imposed .", "twist is not considered , i.e. , @xmath32 .", "of course , the analysis can be straightforwardly extended to the case of different splay and bend constants @xmath4 and @xmath6 , respectively , but this aspect of the problem is important only if the analysis is supposed to yield quantitative results for a specific nematic liquid crystal .", "the surface contribution @xmath33 $ ] includes the anchoring energy for which the rapini - papoular form [ eq .  ( [ eq3 ] ) ] is adopted :    .", "the upper flat substrate induces strong anchoring , i.e. , @xmath31 , while the lower substrate is characterized by a surface pattern of period @xmath24 in @xmath25 direction .", "the system is translational invariant in @xmath26 direction perpendicular to the plane of the figure . in ( a ) the lower sinusoidally grating surface ( groove depth @xmath34 )", "is endowed with an alternating stripe pattern of locally homeotropic anchoring ( white bars ) and homogeneous planar anchoring ( black bars ) .", "the period of the chemical pattern is half the period of the surface grating @xmath24 . in ( c ) and", "( d ) a pure geometrically structured lower substrate and pure chemically patterned lower substrate , respectively , are shown .", "the anchoring direction at the substrates is schematically represented by black rods .", "quantitatively reliable predictions of the phase behavior of a nematic liquid crystal confined between two substrates at an _ arbitrary _ mean distance @xmath35 in ( a ) can be achieved if the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] is analyzed . to this end", "the free energy functional of the nematic liquid crystal confined between two substrates at a _ single _ and rather small mean distance @xmath36 in ( b ) has to be minimized as is discussed in the main text.,width=302 ]    @xmath37=}\\nonumber \\\\&&\\!\\!\\!\\!\\!\\frac{l}{2}\\int^p_0 dx\\ , w(x)\\frac{\\left(-\\sin(\\theta_0(x))z'_0(x)+\\cos(\\theta_0(x))\\right)^2 } { \\sqrt{1+(z'_0(x))^2}},\\end{aligned}\\ ] ]    where @xmath38 .", "equations ( [ eq4 ] ) and ( [ eq5 ] ) together with the boundary condition for the nematic director at the upper surface completely specify the free energy functional for the system under consideration .", "the euler - langrange equations resulting from the stationary conditions of the total free energy with respect to the nematic director field can be solved numerically on a sufficiently fine two - dimensional grid using an iterative method . in order to obtain both stable and metastable configurations , different types of initial configurations", "however , due to the pattern of the lower surface the determination of the director field and the phase diagram turns out to be a challenging numerical problem in particular in the case of large cell widths @xmath35 .", "moreover , the energy barrier between two metastable states can not be determined this way .      here", "we map the free energy functional @xmath39 $ ] [ eq .", "( [ eq4 ] ) ] of a nematic liquid crystal cell with _ arbitrary _", "width @xmath35 and arbitrary anchoring angle @xmath40 at the upper surface ( see fig .", "[ fig1 ]  ( a ) ) onto the effective free energy function @xmath41 where the average surface director orientation @xmath42 @xcite at the lower patterned surface is given by @xmath43 the effective surface free energy function @xmath44 characterizing the anchoring energy at the patterned surface can be written as @xmath45 in order to calculate @xmath42 and @xmath44 explicitly , we first determine numerically the minimum of the free energy @xmath46 of the nematic liquid crystal cell for a _ single _ and rather small value @xmath47 and arbitrary anchoring angle @xmath48 at the upper surface ( see fig .", "[ fig1 ]  ( b ) ) .", "thereafter the phase behavior , energy barriers between metastable states , and effective anchoring angles for the system of interest ( fig .", "[ fig1 ]  ( a ) ) can be obtained for _ arbitrary _ values of @xmath35 and @xmath40 from @xmath49 as a function of the single variable @xmath42 .", "such a calculation is considerably less challenging than minimizing the original free energy functional @xmath39 $ ] with respect to @xmath50 on a two - dimensional @xmath51 grid .", "however , the effective free energy method is applicable only if eq .", "( [ eq7 ] ) can be inverted in order to obtain @xmath52 which is needed as input into eq .", "( [ eq8 ] ) .", "the condition for this inversion follows from eq .", "( [ eq7 ] ) : @xmath53 moreover , @xmath54 is practically independent of the cell width @xmath36 provided @xmath55 implying that the interfacial region above the lower substrate does not extend to the upper substrate .    before studying the nematic liquid crystal in contact with the patterned substrates shown in fig .  [ fig1 ]", "it is instructive to analyze first the nematic liquid crystal confined between two homogeneous flat substrates at a distance @xmath35 .", "the upper surface induces strong anchoring , i.e. , .", "the free energy functional defined in eq .", "( [ eq4 ] ) follows as @xmath56&\\!\\!=\\!\\!&\\frac{klp}{2}\\int\\limits_0^d dz\\ , \\left(\\frac{d \\theta(z)}{d z}\\right)^2+f_s(\\theta_0).\\end{aligned}\\ ] ] the solution of the euler - langrange equation @xmath57 subject to the boundary condition at the upper surface @xmath58 interpolates linearly between the top and bottom surfaces :    @xmath59    where @xmath60 follows from the boundary condition at the lower surface . with this solution of the euler - lagrange equation", "the minimized free energy function reads @xmath61 it follows directly from eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) that @xmath62 , @xmath63 , and @xmath64 hence in the case of homogeneous confining substrates the effective free energy function [ eq .", "( [ eq13 ] ) ] agrees exactly with the minimized free energy function of the original system [ eq .", "( [ eq12 ] ) ] .", "in the previous section we have shown that we can describe a nematic liquid crystal confined between two homogeneous substrates by the effective free energy function [ eq .  ( [ eq6 ] ) ] . in this subsection", "we apply this approach to the particular case of a nematic liquid crystal confined between a chemically patterned sinusoidal surface and a flat substrate with strong homeotropic anchoring ( see fig .", "[ fig1 ]  ( a ) ) . the surface profile of the grating surface", "is given by @xmath65 , where @xmath34 is the groove depth and @xmath66 is the period .", "as figure [ fig1 ] ( a ) illustrates , the surface exhibits a pattern consisting of alternating stripes with locally homeotropic and homogeneous planar anchoring .", "the projection of the widths of the stripes onto the @xmath25 axis is @xmath67 and the anchoring strength is specified by a periodic step function : @xmath68 and @xmath69 for values of @xmath25 on the homeotropic and planar stripes , respectively .", "figure [ fig2 ] ( a ) displays @xmath70 ( dashed line ) and @xmath71 ( solid line ) for @xmath72 , @xmath73 , and @xmath74 . the shapes of @xmath70 as a function of @xmath48 and @xmath71 as a function of @xmath42 are rather similar because @xmath75 ( see eq .", "( [ eq8 ] ) ) for this set of model parameters .", "figure [ fig2 ] ( b ) displays the phase diagram plotted as a function of the anchoring angle @xmath40 at the upper substrate and the mean separation of the substrates @xmath35 .", "the calculations demonstrate the existence of two ( stable or metastable ) nematic director configurations : the homeotropic ( h ) phase , in which the director field is almost uniform and parallel to the anchoring direction imposed at the upper surface , i.e. , @xmath76 , and the hybrid aligned nematic ( han ) phase , in which the director field varies from @xmath77 at the upper surface to nearly planar orientation through the cell .", "note that there are two han textures : han@xmath78 and han@xmath79 corresponding to positive and negative average surface angles at the lower surface ( see also fig .  [", "fig : diagram ] below ) . for small anchoring angles @xmath40 the han phases", "are stable provided the cell width is larger than @xmath80 ( more precisely , the han@xmath78 texture is stable for @xmath81 while the han@xmath79 texture is stable for @xmath82 and they coexist at @xmath83 ) . for smaller distances between the substrates", "@xmath84 the han phases are no longer stable because distortions of the director field are too costly in the presence of the dominating strong anchoring at the upper surface . the comparison of the phase boundary of thermal equilibrium as obtained from the effective free energy method [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) , and solid line in fig .  [ fig2 ]  ( b ) ] and the direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) , and diamonds in fig .  [ fig2 ]  ( b ) ] demonstrate the reliability of the effective free energy method .", "( see fig .", "[ fig1 ]  ( b ) ) .", "@xmath28 is the extension of the cell in the invariant @xmath26 direction , @xmath85 is the isotropic elastic constant , and the anchoring strength on the homeotropic stripes ( white bars in fig .", "[ fig1 ]  ( b ) ) and planar anchoring stripes ( black bars in fig .", "[ fig1 ]  ( b ) ) are @xmath72 and @xmath73 , respectively .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 ( see fig .  [ fig1 ]  ( a ) ) .", "the solid line denotes first order phase transitions between a homeotropic ( h ) and hybrid aligned nematic ( han@xmath78 and han@xmath79 ) phases . at @xmath86 and @xmath87 there is a triple point where the han@xmath78 , han@xmath79 , and h states coexist .", "the solid circle marks the critical point at @xmath88 and @xmath89 .", "the limits of metastability of the han@xmath78 ( 1 ) and the h ( 2 ) state are denoted by the dot - dashed lines .", "the limit of metastability of the h state for @xmath82 and the han@xmath79 state are not shown for clearness .", "the lines and the solid circle follow from analyzing the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] while the diamonds represent the phase boundary of thermal equilibrium as obtained from a direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5])].,width=302 ]     ( see fig .  [ fig1 ]  ( b ) ) . the anchoring strength on the homeotropic stripes and planar anchoring stripes are @xmath90 and @xmath91 , respectively .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 ( see fig .", "[ fig1 ]  ( a ) ) .", "the solid line denotes the first order phase transition between a homogeneous ( h ) and hybrid aligned nematic ( han@xmath78 ) phase .", "the solid circle marks the critical point at @xmath92 and @xmath93 .", "the limits of metastability of the han@xmath78 ( 1 ) and the h ( 2 ) state are denoted by the dot - dashed lines .", "the lines follow from analyzing the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] while the diamonds and the solid circle represent the phase boundary and a critical point as obtained from a direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) ] . for @xmath94", "the phase transition as well as the limits of metastability can not be determined using the effective free energy function because @xmath95 is not known in the region close to its maximum . for clearness ,", "only the phase diagram for positive @xmath40 is shown ( c.f .", "[ fig2]).,width=302 ]    we note that the phase transition between the h and han textures is first order despite the fact that the effective surface free energy favors monostable planar anchoring , i.e. , @xmath44 exhibits only a minimum at @xmath96 in the interval @xmath97 $ ] .", "a first order phase transition in a nematic liquid crystal device with a monostable anchoring condition on a homogeneous lower substrate has been predicted for the special case @xmath98 in refs .", "@xcite using the empirical expression @xmath99 as input into eq .", "( [ eq13 ] ) . for @xmath100", "this surface free energy has two minima at @xmath101 and @xmath102 in the interval @xmath103 $ ] , and as such is bistable , while for @xmath104 , only the minimum @xmath101 exists , i.e. , the surface is monostable . to study the stability limit of the h phase", "we expand @xmath105 in eq .", "( [ eq13 ] ) around @xmath101 up to sixth order : @xmath106 the h phase corresponds to a local minimum of @xmath105 in eq .", "( [ eq13 ] ) if @xmath107 , where @xmath108 a standard bifurcation analysis reveals that the transition from the h phase to the han phase can be either first order or continuous .", "the transition is continuous if @xmath109 , first order if @xmath110 , and @xmath111 corresponds to a tricritical point . in the case of a first order phase transition", "the phase boundary of thermal equilibrium is given by @xmath112 we emphasize that the order of the phase transition depends only on the surface free energy @xmath113 close to @xmath101 for @xmath98 .", "therefore it is possible to have a first order phase transition even with a monostable surface characterized by a monotonic surface free energy such as the one shown in fig .", "[ fig2 ]  ( a ) for @xmath114 $ ] or the empirical equation ( [ eq14 ] ) with @xmath104 as well the more general expression @xcite @xmath115\\end{aligned}\\ ] ] with appropriate parameters @xmath116 and @xmath117 .", "first order phase transitions between the h and han texture are of particular interest for bistable liquid crystal displays . in a bistable liquid crystal display", "the two molecular configurations corresponding to light and dark states are locally stable in the thermodynamic space when the applied voltage is removed @xcite . therefore , power is needed only to switch from one stable state to another , in contrast to monostable liquid crystal displays which require power to switch and to maintain the light and the dark states .", "we now turn our attention to the case that it is not possible to evaluate @xmath118 from @xmath119 [ eq .  ( [ eq7 ] ) ] because the condition for this inversion is not satisfied [ eq .  ( [ eq9 ] ) ] . to this end", "we have chosen the parameters @xmath90 , @xmath91 , and @xmath74 for the system shown in fig .", "[ fig1 ]  ( a ) .", "figure [ fig3 ] ( a ) display ( dashed line ) and ( solid line ) while the corresponding phase diagram is shown in fig .", "[ fig3 ]  ( b ) . as is apparent from the solid line in fig .", "[ fig3 ]  ( a ) it is not possible to determine the effective surface free energy function for the all values of @xmath42 because the upper flat substrate at @xmath36 is too far away from the lower patterned substrate in order to induce all possible average anchoring orientations @xmath42 . in other words , the anchoring energy at the patterned substrate is too large to be balanced by the elastic energy for the chosen mean distance @xmath36 between the substrates ( see fig .", "[ fig1 ]  ( b ) ) . nevertheless fig .", "[ fig3 ]  ( b ) demonstrates that even this partial information about the effective surface free energy function can be used to calculate the phase diagram for cell widths sufficiently larger than the width at the critical point @xmath120 .      in the last subsection we have shown that with a suitable chemical and geometrical surface morphology on one of the interior surfaces of a liquid crystal cell", ", two stable nematic director configurations can be supported .", "the zenithally bistable nematic devices that have been studied recently @xcite consist of a nematic liquid crystal confined between a chemically homogeneous grating surface ( see fig .", "[ fig1 ]  ( c ) ) and a flat substrate with strong homeotropic anchoring .", "the profile of the asymmetric surface grating is given by @xmath121 where @xmath34 is the groove depth , @xmath66 the period , and @xmath122 is the `` blazing '' parameter describing the asymmetry of the surface profile", ". such a grating surface has been studied by brown _", "@xcite who found a first order transition between the han state , characterized by a low pretilt angle ( @xmath123 ) , and the h state , characterized by a high pretilt angle ( @xmath124 ) . strictly speaking", ", the h state does not correspond to the homeotropic texture ( see fig .", "[ fig5 ]  ( b ) below ) , but we keep the same notation as in the previous section for consistency . here", "we study phase transitions of a nematic liquid crystal in contact with the blazed surface in a more detail using the effective free energy method discussed in sec .", "ii  c.    in fig .", "[ fig : blazed ]  ( a ) the minimized free energy @xmath70 ( dashed line ) and the calculated effective surface free energy @xmath71 ( solid line ) are shown for the anchoring strength @xmath125 on the grating surface , the groove depth @xmath126 , and the blazing parameter @xmath127 .", "the effective surface free energy is asymmetric with respect to @xmath128 because of the asymmetry of the grating surface . as a consequence", "also the phase diagram , plotted as a function of the anchoring angle on the upper surface @xmath40 and the distance @xmath129 is asymmetric ( fig .  [", "fig : blazed ]  ( b ) ) .", "however , the topology of the phase diagram is the same as in the case of a symmetric substrate ( see figs .", "[ fig2 ] ( b ) , [ fig3 ] ( b ) , and for a more general discussion sec .", "iii  c below ) .", "( c ) and eq .", "( [ eq : blazed ] ) ) .", "the locally homeotropic anchoring strength on the blazed surface is @xmath130 , the groove depth is @xmath131 and the blazing parameter is @xmath127 .", "@xmath28 is the extension of the cell in the invariant @xmath26 direction and @xmath85 is the isotropic elastic constant .", "the effective surface free energy ( as well as ) is periodic with the period @xmath132 but it is asymmetric with respect to @xmath133 .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 with the same line code as in figs .", "[ fig2 ] and [ fig3 ] .", "the triple point ( where the han@xmath78 , han@xmath79 and h phases coexist ) is at @xmath134 and @xmath135 , and the critical points ( solid circle ) are at @xmath136 , @xmath137 and @xmath138 , @xmath139", ". for clearness , only the limits of metastability of the han@xmath79 ( 1 ) phase and the h ( 2 ) phase ( for negative @xmath40 ) are shown.,width=302 ]    ( c ) and eq .", "( [ eq : blazed ] ) ) with local homeotropic anchoring and a flat surface with strong homeotropic anchoring ( @xmath98 ) as a function of the groove depth @xmath140 and the cell width @xmath129 .", "the anchoring strength on the blazed surface is @xmath130 .", "the lines correspond to different values of @xmath122 and denote first order transitions between homeotropic ( h ) and hybrid aligned ( han@xmath79 ) phases . for small @xmath140 , the lines extend to @xmath141 corresponding to a first order _ anchoring _ transition between planar and homeotropic phases . upon increasing", "@xmath140 the first order transition lines end at critical points which are not shown in the figure .", "( b ) the energy barrier at the first order transitions with the same line code as in ( a ) .", "the lines have been obtained from the total effective free energy ( see eq .", "( [ eq6 ] ) ) , while the diamonds correspond to the energy barriers between the planar and homeotropic effective anchoring which follow from considering only the surface contribution @xmath44.,width=302 ]    ) ) for the hybrid aligned ( han@xmath79 ) phase in ( a ) and the homeotropic ( h ) phase in ( b ) on the lines of the first order transitions ( see fig .  [ fig4 ] ) for a nematic liquid crystal confined between the blazed surface ( see fig .  [ fig1 ]  ( c ) and eq .", "( [ eq : blazed ] ) ) with local homeotropic anchoring and a flat surface with strong homeotropic anchoring ( @xmath98 ) .", "the diamonds in ( a ) and ( b ) correspond to the two minima of the effective surface energy function @xmath142 .", "the model parameters and the line code are the same as in fig .", ", width=302 ]    we now concentrate on the most interesting ( from a practical point of view ) case of strong homeotropic anchoring ( @xmath86 ) at the upper homogeneous surface .", "figure [ fig4 ] ( a ) displays the phase diagram for a few values of the blazing parameter @xmath122 and a fixed value of the local homeotropic anchoring strength on the grating surface @xmath130 . for a fixed cell width @xmath129 , asymmetry ( @xmath143 ) leads to a decrease of the groove depth @xmath140 at which there is a first order transition between the han and the h phases , as compared to the nematic liquid crystal cell with the symmetric surface grating ( @xmath144 ) . upon increasing the groove depth", "@xmath140 the transition line ends at a critical point ( not shown in the figure ) , while it diverges as @xmath145 .", "the groove depth @xmath146 corresponds to an _ anchoring _ ( or surface ) transition between low tilt and high tilt _ surface _", "states which are the homeotropic and planar effective anchoring states , respectively , in the case @xmath144 .", "the effective free energy method allows one to calculate an energy barrier @xmath147 between the two bistable states , which is _ not _", "feasible by the direct numerical minimization of the free energy functional .", "the results of our calculations are shown in fig .", "[ fig4 ]  ( b ) .", "the asymmetry of the surface grating leads to a decrease of the energy barrier . with increasing groove depth @xmath140 the energy barrier decreases and eventually vanishes upon approaching the critical point .", "the diamonds in fig .", "[ fig4 ]  ( b ) denote the energy barriers for the abovementioned anchoring transitions of a nematic liquid crystal in contact with a single grating surface .", "the energy barrier between two bistable states is an important quantity for the design of a zenithally bistable nematic device .", "too small energy barrier , as compared to @xmath148 , would cause spontaneous switching between the two states because of thermal fluctuations , while enlarging the energy barrier leads to an increase of the power consumption . using the calculated values of @xmath147 ( see fig .", "[ fig4 ]  ( b ) ) one can estimate the energy barrier in a real nematic liquid crystal cell .", "for instance , for a cell of area @xmath149 and of width @xmath150 , and taking the typical values @xmath151 and @xmath152 , one obtains @xmath153 for @xmath154 and @xmath155 , which seems to be an acceptable value .", "another important quantity in zenithally bistable nematic devices is the average director orientation at the grating surface in the two degenerate states .", "the average surface director in the han ( @xmath156 ) and h ( @xmath157 ) states is shown in fig .", "[ fig5 ] for the same model parameters as in fig .", "[ fig4 ] and for the values of @xmath140 and @xmath129 corresponding to the coexistence line .", "the asymmetry of the surface grating leads to a decrease of @xmath156 and an increase of @xmath158 . for a fixed value of @xmath122", ", the difference between the two angles decreases with increasing the groove depth @xmath140 and finally vanishes upon approaching the critical point ( not shown in the figure ) .", "hence the asymmetry of the surface grating leads to a decrease of the groove depth at which the bistability is observed , which improves optical properties @xcite , and to a decrease of the energy barrier , which lowers the power consumption of a zenithally bistable nematic device . on the other hand ,", "also the difference between the two bistable states decreases which impairs optical properties of such a device .", "in section [ iiia ] we have discussed phase diagrams in the @xmath159 plane which can be described in terms of the surface free energy given by eq .  .", "however , it is instructive to consider a more general situation that the surface free energy follows from a truncation of the fourier expansion given in eq .  .", "to be able to study both symmetric and asymmetric surfaces we assume a natural generalization of eq .", ", namely @xmath160\\,,\\end{aligned}\\ ] ] which reduces to eq .   in the case of a symmetric surface characterized by @xmath161 .", "the angle @xmath162 that minimizes @xmath163 ( see eq .  ) is a function of @xmath40 and @xmath35 .", "the derivative @xmath164 is the susceptibility of the system that diverges at the critical thickness @xmath165 and remains finite and positive for @xmath107 . from eq .   the conditions for the critical angle @xmath166 follow as @xmath167 implying that @xmath120 and @xmath166 depend only on the form of @xmath113 .", "plane for the surface free energy given by eq . .", "@xmath168 and @xmath169 denote different ( usually non - uniform ) textures , the thick lines correspond to first order transitions , and black circles mark critical points .", "when @xmath11 ( defined on the unit circle @xmath170 ) has one minimum and one maximum the phase diagram can be either of type ( a ) or ( b ) .", "when @xmath11 has two minima and two maxima the phase diagram is of type ( b ) unless the minima are of equal depth , in which case it is of type ( c ) .", "note that the lines @xmath171 are identified with each other and the phase diagram can be considered as being on a cylindrical surface.,scaledwidth=49.0% ]    the extremes of @xmath11 given by eq .", "can be found easily only in the case of symmetric or antisymmetric ( @xmath172 ) surface and the same concerns the position of critical point . in this work , however , we are interested rather in possible topologies of the phase diagram in the @xmath159 plane , which result from eqs . and , and not in the exact location of critical points or transition lines . to draw schematic phase diagrams we consider @xmath11 as a function defined on the unit circle @xmath170", "depending on the parameters @xmath116 and @xmath117 , @xmath113 has either one minimum and one maximum or two minima and two maxima .", "this conclusion applies also to the function @xmath173 , thus , there can be either one or two critical points in the phase diagram ( see eqs . and ) . in the limit of large @xmath35", ", there is always a first order phase transition between two non - uniform textures corresponding to the opposite orientations of the director at @xmath21 .", "this is because @xmath174 is not a periodic function of @xmath175 at fixed @xmath40 . to find @xmath40 at the transition we expand @xmath11 around its deepest minimum ( denoted @xmath176 ) , which leads to the approximate free energy : @xmath177 where @xmath178 is the extrapolation length .", "since @xmath176 and @xmath179 are equivalent minima of @xmath11 , and both @xmath176 and @xmath40 are allowed to vary in the interval @xmath180 $ ] , the transition occurs at @xmath181 if @xmath182 or at @xmath183 if @xmath184 . with the above information we can now draw schematically the phase diagram ( see fig .  [ fig : diagram ] ) .", "since the vertical lines at @xmath171 are to be identified with each other the phase diagram can be considered on a cylindrical surface .", "away from the transition lines there is a smooth evolution from one texture to another .", "this means that at fixed @xmath35 it is possible to transform smoothly the @xmath185 texture into the @xmath186 texture , i.e. , without crossing the transition line , even for @xmath187 .", "we note that in some range of parameters , the phase diagram for @xmath11 with one minimum is topologically indistinguishable from that for @xmath11 with two minima ( fig .", "[ fig : diagram ]  ( b ) ) ; in both cases there are two critical points and a triple point .", "if @xmath11 has two equal minima ( e.g. , at @xmath101 and @xmath188 in the case of symmetric surface ) the triple point disappears and the first order transition lines extend to @xmath189 , as shown in fig .", "[ fig : diagram ]  ( c ) .", "we have studied the phase behavior of a nematic liquid crystal confined between a flat and a patterned substrate ( fig .", "[ fig1 ] ) using the frank - oseen model [ eq .", "( [ eq4 ] ) ] and the rapini - papoular surface free energy [ eq .", "( [ eq5 ] ) ] . an expression for the effective free energy function of the system [ eq .", "( [ eq6 ] ) ] was derived by determining an effective surface free energy characterizing the anchoring energy at the patterned surface [ eq .  ( [ eq8 ] ) ] . using the effective free energy function ,", "we have determined the phase behavior of the nematic liquid crystal confined between a flat surface with strong anchoring and a chemically patterned sinusoidal surface ( fig .", "[ fig1 ] ( a ) ) , finding first order transitions between a homeotropic texture ( h ) and hybrid aligned nematic ( han ) textures ( figs .", "[ fig2 ] ( b ) and [ fig3 ] ( b ) ) .", "it is possible to have a first order phase transition even with a monostable surface characterized by a monotonic surface free energy function ( fig .", "[ fig2 ]  ( a ) , @xmath114 $ ] ) .", "in addition we have performed direct minimizations of the original free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) ] on a two - dimensional grid and found remarkably good agreement with the phase boundaries resulting from the effective energy function analysis ( figs", ".  [ fig2 ] ( b ) and [ fig3 ] ( b ) ) .", "hence quantitatively reliable predictions of the phase behavior can be achieved using the effective free energy method .    using this method ,", "we have also studied the phase behavior ( fig .", "[ fig : blazed ] ( b ) ) of a nematic liquid crystal confined between a chemically uniform , asymmetrically grooved substrate ( fig .", "[ fig1 ]  ( c ) and eq .", "( [ eq : blazed ] ) ) with locally homeotropic anchoring and a flat substrate with strong homeotropic anchoring , which is a typical setup for a zenithally bistable nematic device @xcite .", "the asymmetry of the grating substrate leads to a decrease of the groove depth at which a first order transition between the h and han phases occurs ( fig .", "[ fig4 ]  ( a ) ) .", "moreover , we have determined the energy barrier between the two coexisting states ( fig .", "[ fig4 ]  ( b ) ) .", "our calculations show that the energy barrier decreases with increasing the asymmetry of the grating surface but it is well above @xmath190 for a typical nematic liquid crystal cell .", "in addition , the average director orientation at the grating surface in two bistable states has been calculated ( fig .", "[ fig5 ] ) .", "the difference between the two bistable states vanishes with increasing substrate asymmetry , which has a negative effect on the optical properties of a zenithally bistable nematic device .", "we have also generalized the model of the effective surface free energy considered by parry - jones _", "@xcite to the case of asymmetric structured substrates and obtained three possible types of the phase diagram in the plane spanned by the orientation of the director at the homogeneous surface and the thickness of the nematic cell .", "the asymmetry of the substrate causes only a shift of transition lines and critical points , compared to the symmetric case , but does not change the topology of the phase diagram .", "finally , we have verified that this model allows one to reproduce qualitatively the phase diagram of a nematic liquid crystal confined between a homogeneous planar substrate and an asymmetrically grooved surface ( fig .", "[ fig : blazed ]  ( b ) and fig .", "[ fig : diagram ]  ( b ) ) ."], "abstract_text": ["<S> we study the phase behavior of a nematic liquid crystal confined between a flat substrate with strong anchoring and a patterned substrate whose structure and local anchoring strength we vary . by first evaluating an effective surface free energy function characterizing the patterned substrate we derive an expression for the effective free energy of the confined nematic liquid crystal . </S>", "<S> then we determine phase diagrams involving a homogeneous state in which the nematic director is almost uniform and a hybrid aligned nematic state in which the orientation of the director varies through the cell . </S>", "<S> direct minimization of the free energy functional were performed in order to test the predictions of the effective free energy method . </S>", "<S> we find remarkably good agreement between the phase boundaries calculated from the two approaches . </S>", "<S> in addition the effective energy method allows one to determine the energy barriers between two states in a bistable nematic device . </S>"], "labels": null, "section_names": ["introduction", "effective free energy function", "applications", "summary"], "sections": [["the interest in anchoring phenomena and phenomena in confined nematic liquid crystals has largely been driven by their potential use in liquid crystal display devices .", "the twisted nematic liquid crystal cell serves as an example .", "it consists of a nematic liquid crystal confined between two parallel walls , both providing homogeneous planar anchoring but with mutually perpendicular easy directions . in this case", "the orientation of the nematic director is tuned by the application of an external electric or magnetic field .", "a precise control of the surface alignment extending over large areas is decisive for the functioning of such devices .", "most studies have focused on nematic liquid crystals in contact with laterally uniform substrates . on the other hand substrate inhomogeneities", "arise rather naturally as a result of surface treatments such as rubbing .", "thus the nematic texture near the surface is in fact non - uniform .", "this non - uniformity , however , is smeared out beyond a decay length proportional to the periodicity of the surface pattern .", "very often the thickness of the non - uniform surface layer is considerably smaller than both the wavelength of visible light and the thickness of the nematic cell , i.e. , the distance between the two confining parallel walls", ". hence optical properties of the nematic liquid crystal confined between such substrates correspond to those resulting from effective , uniform substrates .", "more recent developments have demonstrated that surfaces patterned with a large periodicity of some micrometers are of considerable interest from a technological point of view ( see , e.g. , ref .", "@xcite and references therein ) .", "a new generation of electro - optical devices relies on nematic liquid crystals with patterned orientation of the nematic director over small areas which can be achieved by chemically patterning the confining surfaces .", "for example , to produce flat - panel displays with wide viewing angles one can use pixels that are divided into sub - pixels , where each sub - pixel is defined by a different orientation of the nematic director , which is induced by the surface structure and subsequently tuned by the electric field .", "in addition to the technological relevance , nematic liquid crystals in contact with non - uniform substrates provide the opportunity to study basic phenomena such as effective elastic forces between the substrates and phase transitions between various competing nematic textures ( see , e.g. , ref .", "@xcite and references therein ) .    whereas the influence of homogeneous confining substrates on nematic liquid crystals is now well understood ,", "the phase behavior of nematic liquid crystals in contact with chemically or geometrically patterned substrates is still debated .", "one might suppose that theoretical calculations based on continuum theories should resolve the properties of nematic liquid crystals in contact with patterned substrates @xcite .", "however , such calculations are numerically demanding because two- or three - dimensional grids have to be used because of the broken symmetry due to the surface pattern .", "moreover , it is very challenging to determine metastable states and energy barriers between them which are important for the understanding of bistable nematic devices @xcite . in the present paper", "we adopt a different strategy which takes the advantage of the finite decay length characterizing the influence of the surface pattern on the nematic liquid crystal in the direction perpendicular to the substrate .", "we determine first an anchoring energy function and an average surface director orientation of the patterned substrate and obtain an effective free energy for the nematic liquid crystal cell under consideration .", "we find remarkably good agreement between the phase diagrams of various systems calculated using this effective free energy function on the one hand and the original free energy functional on the other hand ."], ["the continuum theory for liquid crystals has its origin dating back to at least the work of oseen @xcite and zocher @xcite .", "this early version of the continuum theory for nematic liquid crystals played an important role for the further development of the static theory and its more direct formulation by frank @xcite .", "the frank theory is formulated in terms of the so - called nematic director @xmath0 , @xmath1 , and its possible spatial distortions .", "the nematic director describes the direction of the locally averaged molecular alignment in liquid crystals . in a nematic liquid crystal", "the centers of mass of the liquid crystal molecules do not exhibit long - ranged order .", "the molecules can translate freely while being aligned , on average , parallel to one another and to the nematic director .", "it is known that if an initially uniform nematic liquid crystal is distorted by external forces , it relaxes back to the uniform state after the disturbing influence is switched off , signaling that the uniform configuration represents a thermodynamically stable state .", "therefore it is assumed that there is a cost in free energy associated with elastic distortions of the nematic director of the form @xmath2&=&\\frac{1}{2}\\int_v", "d^3 r\\,\\left[k_{11}\\left(\\nabla\\cdot\\hat{{\\bf n}}\\right)^2\\right.\\nonumber \\\\&+&\\!\\!\\!\\left.k_{22}(\\hat{{\\bf n}}\\cdot(\\nabla \\times\\hat{{\\bf n}}))^2 + k_{33}(\\hat{{\\bf n}}\\times(\\nabla \\times \\hat{{\\bf n}}))^2\\right],\\nonumber \\\\ & & \\label{eq1}\\end{aligned}\\ ] ] where @xmath3 is the volume accessible to the nematic liquid crystal and @xmath4 , @xmath5 , and @xmath6 are elastic constants associated with splay , twist , and bend distortions , respectively .", "the elastic constants depend on temperature and are commonly of the order @xmath7 to . sometimes , for example , when the relative values of the elastic constants are unknown or when the resulting euler - lagrange equations are complicated , the one - constant approximation @xmath8 is made . in this case", "the elastic free energy functional reduces to @xmath9&=&\\frac{k}{2}\\int_v d^3 r\\,\\left(\\nabla\\hat{{\\bf n}}\\right)^2\\,.\\end{aligned}\\ ] ] in the presence of surfaces the bulk free energy @xmath10 must be supplemented by the surface free energy @xmath11 such that the total free energy is given by @xmath12 . in the corresponding equilibrium euler - lagrange equations @xmath13 , @xmath11 leads to appropriate boundary conditions .", "the description of the nematic director alignment at the surfaces forming the boundaries is called anchoring .", "in addition to the so - called free boundary condition where there is no anchoring , one considers weak and strong anchoring . if there are no anchoring conditions imposed on @xmath14 at the boundary , the bulk free energy @xmath15 is minimized using standard techniques of the calculus of variations . in the case of strong anchoring it", "is also sufficient to minimize the bulk free energy but subject to @xmath14 taking prescribed fixed values at the boundary . in the case of weak", "anchoring the total free energy @xmath16 , which includes the surface free energy @xmath11 , has to be minimized .", "the most commonly used expression for the surface free energy is of the form proposed by rapini and papoular by @xcite : @xmath17=\\frac{1}{2}\\int_{s } d^2 r\\,w({\\bf r } ) ( \\hat{{\\bf n}}\\cdot{\\nu})^2.\\ ] ] the integral runs over the boundary and @xmath18 is the corresponding anchoring strength that characterizes the surface .", "the local unit vector perpendicular to the surface is denoted as @xmath14 . for negative @xmath19", ", this contribution favors an orientation of the molecules perpendicular to the surface , while positive @xmath19 favor degenerate planar orientations at the surface .", "the absolute value of the anchoring strength is commonly of the order @xmath20 to .", "here we consider a nematic liquid crystal confined between a patterned substrate at @xmath21 and a flat substrate at @xmath22 where the @xmath23 axis is normal to the flat substrate . as figs .", "[ fig1 ] ( a ) , ( c ) , and ( d ) illustrate , the lower substrate is characterized by geometrical and/or chemical patterns of periodicity @xmath24 along the @xmath25 axis . moreover , the system is translationally invariant in the @xmath26 direction . within the one - constant approximation [ eq .  ( [ eq2 ] ) ] the total free energy functional is given by @xmath27=}\\nonumber \\\\&&\\frac{kl}{2}\\int\\limits_0^p dx\\ , \\int\\limits_{z_0(x)}^d dz\\ , [ \\nabla\\theta(x , z)]^2+f_s[\\theta(x , z_0(x))]\\ , , \\label{eq4}\\end{aligned}\\ ] ] where @xmath28 is the extension of the system in @xmath26 direction , @xmath29 is the surface profile of the patterned substrate , and @xmath30 . at the upper surface", "strong anchoring @xmath31 is imposed .", "twist is not considered , i.e. , @xmath32 .", "of course , the analysis can be straightforwardly extended to the case of different splay and bend constants @xmath4 and @xmath6 , respectively , but this aspect of the problem is important only if the analysis is supposed to yield quantitative results for a specific nematic liquid crystal .", "the surface contribution @xmath33 $ ] includes the anchoring energy for which the rapini - papoular form [ eq .  ( [ eq3 ] ) ] is adopted :    .", "the upper flat substrate induces strong anchoring , i.e. , @xmath31 , while the lower substrate is characterized by a surface pattern of period @xmath24 in @xmath25 direction .", "the system is translational invariant in @xmath26 direction perpendicular to the plane of the figure . in ( a ) the lower sinusoidally grating surface ( groove depth @xmath34 )", "is endowed with an alternating stripe pattern of locally homeotropic anchoring ( white bars ) and homogeneous planar anchoring ( black bars ) .", "the period of the chemical pattern is half the period of the surface grating @xmath24 . in ( c ) and", "( d ) a pure geometrically structured lower substrate and pure chemically patterned lower substrate , respectively , are shown .", "the anchoring direction at the substrates is schematically represented by black rods .", "quantitatively reliable predictions of the phase behavior of a nematic liquid crystal confined between two substrates at an _ arbitrary _ mean distance @xmath35 in ( a ) can be achieved if the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] is analyzed . to this end", "the free energy functional of the nematic liquid crystal confined between two substrates at a _ single _ and rather small mean distance @xmath36 in ( b ) has to be minimized as is discussed in the main text.,width=302 ]    @xmath37=}\\nonumber \\\\&&\\!\\!\\!\\!\\!\\frac{l}{2}\\int^p_0 dx\\ , w(x)\\frac{\\left(-\\sin(\\theta_0(x))z'_0(x)+\\cos(\\theta_0(x))\\right)^2 } { \\sqrt{1+(z'_0(x))^2}},\\end{aligned}\\ ] ]    where @xmath38 .", "equations ( [ eq4 ] ) and ( [ eq5 ] ) together with the boundary condition for the nematic director at the upper surface completely specify the free energy functional for the system under consideration .", "the euler - langrange equations resulting from the stationary conditions of the total free energy with respect to the nematic director field can be solved numerically on a sufficiently fine two - dimensional grid using an iterative method . in order to obtain both stable and metastable configurations , different types of initial configurations", "however , due to the pattern of the lower surface the determination of the director field and the phase diagram turns out to be a challenging numerical problem in particular in the case of large cell widths @xmath35 .", "moreover , the energy barrier between two metastable states can not be determined this way .      here", "we map the free energy functional @xmath39 $ ] [ eq .", "( [ eq4 ] ) ] of a nematic liquid crystal cell with _ arbitrary _", "width @xmath35 and arbitrary anchoring angle @xmath40 at the upper surface ( see fig .", "[ fig1 ]  ( a ) ) onto the effective free energy function @xmath41 where the average surface director orientation @xmath42 @xcite at the lower patterned surface is given by @xmath43 the effective surface free energy function @xmath44 characterizing the anchoring energy at the patterned surface can be written as @xmath45 in order to calculate @xmath42 and @xmath44 explicitly , we first determine numerically the minimum of the free energy @xmath46 of the nematic liquid crystal cell for a _ single _ and rather small value @xmath47 and arbitrary anchoring angle @xmath48 at the upper surface ( see fig .", "[ fig1 ]  ( b ) ) .", "thereafter the phase behavior , energy barriers between metastable states , and effective anchoring angles for the system of interest ( fig .", "[ fig1 ]  ( a ) ) can be obtained for _ arbitrary _ values of @xmath35 and @xmath40 from @xmath49 as a function of the single variable @xmath42 .", "such a calculation is considerably less challenging than minimizing the original free energy functional @xmath39 $ ] with respect to @xmath50 on a two - dimensional @xmath51 grid .", "however , the effective free energy method is applicable only if eq .", "( [ eq7 ] ) can be inverted in order to obtain @xmath52 which is needed as input into eq .", "( [ eq8 ] ) .", "the condition for this inversion follows from eq .", "( [ eq7 ] ) : @xmath53 moreover , @xmath54 is practically independent of the cell width @xmath36 provided @xmath55 implying that the interfacial region above the lower substrate does not extend to the upper substrate .    before studying the nematic liquid crystal in contact with the patterned substrates shown in fig .  [ fig1 ]", "it is instructive to analyze first the nematic liquid crystal confined between two homogeneous flat substrates at a distance @xmath35 .", "the upper surface induces strong anchoring , i.e. , .", "the free energy functional defined in eq .", "( [ eq4 ] ) follows as @xmath56&\\!\\!=\\!\\!&\\frac{klp}{2}\\int\\limits_0^d dz\\ , \\left(\\frac{d \\theta(z)}{d z}\\right)^2+f_s(\\theta_0).\\end{aligned}\\ ] ] the solution of the euler - langrange equation @xmath57 subject to the boundary condition at the upper surface @xmath58 interpolates linearly between the top and bottom surfaces :    @xmath59    where @xmath60 follows from the boundary condition at the lower surface . with this solution of the euler - lagrange equation", "the minimized free energy function reads @xmath61 it follows directly from eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) that @xmath62 , @xmath63 , and @xmath64 hence in the case of homogeneous confining substrates the effective free energy function [ eq .", "( [ eq13 ] ) ] agrees exactly with the minimized free energy function of the original system [ eq .", "( [ eq12 ] ) ] ."], ["in the previous section we have shown that we can describe a nematic liquid crystal confined between two homogeneous substrates by the effective free energy function [ eq .  ( [ eq6 ] ) ] . in this subsection", "we apply this approach to the particular case of a nematic liquid crystal confined between a chemically patterned sinusoidal surface and a flat substrate with strong homeotropic anchoring ( see fig .", "[ fig1 ]  ( a ) ) . the surface profile of the grating surface", "is given by @xmath65 , where @xmath34 is the groove depth and @xmath66 is the period .", "as figure [ fig1 ] ( a ) illustrates , the surface exhibits a pattern consisting of alternating stripes with locally homeotropic and homogeneous planar anchoring .", "the projection of the widths of the stripes onto the @xmath25 axis is @xmath67 and the anchoring strength is specified by a periodic step function : @xmath68 and @xmath69 for values of @xmath25 on the homeotropic and planar stripes , respectively .", "figure [ fig2 ] ( a ) displays @xmath70 ( dashed line ) and @xmath71 ( solid line ) for @xmath72 , @xmath73 , and @xmath74 . the shapes of @xmath70 as a function of @xmath48 and @xmath71 as a function of @xmath42 are rather similar because @xmath75 ( see eq .", "( [ eq8 ] ) ) for this set of model parameters .", "figure [ fig2 ] ( b ) displays the phase diagram plotted as a function of the anchoring angle @xmath40 at the upper substrate and the mean separation of the substrates @xmath35 .", "the calculations demonstrate the existence of two ( stable or metastable ) nematic director configurations : the homeotropic ( h ) phase , in which the director field is almost uniform and parallel to the anchoring direction imposed at the upper surface , i.e. , @xmath76 , and the hybrid aligned nematic ( han ) phase , in which the director field varies from @xmath77 at the upper surface to nearly planar orientation through the cell .", "note that there are two han textures : han@xmath78 and han@xmath79 corresponding to positive and negative average surface angles at the lower surface ( see also fig .  [", "fig : diagram ] below ) . for small anchoring angles @xmath40 the han phases", "are stable provided the cell width is larger than @xmath80 ( more precisely , the han@xmath78 texture is stable for @xmath81 while the han@xmath79 texture is stable for @xmath82 and they coexist at @xmath83 ) . for smaller distances between the substrates", "@xmath84 the han phases are no longer stable because distortions of the director field are too costly in the presence of the dominating strong anchoring at the upper surface . the comparison of the phase boundary of thermal equilibrium as obtained from the effective free energy method [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) , and solid line in fig .  [ fig2 ]  ( b ) ] and the direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) , and diamonds in fig .  [ fig2 ]  ( b ) ] demonstrate the reliability of the effective free energy method .", "( see fig .", "[ fig1 ]  ( b ) ) .", "@xmath28 is the extension of the cell in the invariant @xmath26 direction , @xmath85 is the isotropic elastic constant , and the anchoring strength on the homeotropic stripes ( white bars in fig .", "[ fig1 ]  ( b ) ) and planar anchoring stripes ( black bars in fig .", "[ fig1 ]  ( b ) ) are @xmath72 and @xmath73 , respectively .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 ( see fig .  [ fig1 ]  ( a ) ) .", "the solid line denotes first order phase transitions between a homeotropic ( h ) and hybrid aligned nematic ( han@xmath78 and han@xmath79 ) phases . at @xmath86 and @xmath87 there is a triple point where the han@xmath78 , han@xmath79 , and h states coexist .", "the solid circle marks the critical point at @xmath88 and @xmath89 .", "the limits of metastability of the han@xmath78 ( 1 ) and the h ( 2 ) state are denoted by the dot - dashed lines .", "the limit of metastability of the h state for @xmath82 and the han@xmath79 state are not shown for clearness .", "the lines and the solid circle follow from analyzing the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] while the diamonds represent the phase boundary of thermal equilibrium as obtained from a direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5])].,width=302 ]     ( see fig .  [ fig1 ]  ( b ) ) . the anchoring strength on the homeotropic stripes and planar anchoring stripes are @xmath90 and @xmath91 , respectively .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 ( see fig .", "[ fig1 ]  ( a ) ) .", "the solid line denotes the first order phase transition between a homogeneous ( h ) and hybrid aligned nematic ( han@xmath78 ) phase .", "the solid circle marks the critical point at @xmath92 and @xmath93 .", "the limits of metastability of the han@xmath78 ( 1 ) and the h ( 2 ) state are denoted by the dot - dashed lines .", "the lines follow from analyzing the effective free energy function [ eqs .", "( [ eq6 ] ) - ( [ eq8 ] ) ] while the diamonds and the solid circle represent the phase boundary and a critical point as obtained from a direct minimization of the underlying free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) ] . for @xmath94", "the phase transition as well as the limits of metastability can not be determined using the effective free energy function because @xmath95 is not known in the region close to its maximum . for clearness ,", "only the phase diagram for positive @xmath40 is shown ( c.f .", "[ fig2]).,width=302 ]    we note that the phase transition between the h and han textures is first order despite the fact that the effective surface free energy favors monostable planar anchoring , i.e. , @xmath44 exhibits only a minimum at @xmath96 in the interval @xmath97 $ ] .", "a first order phase transition in a nematic liquid crystal device with a monostable anchoring condition on a homogeneous lower substrate has been predicted for the special case @xmath98 in refs .", "@xcite using the empirical expression @xmath99 as input into eq .", "( [ eq13 ] ) . for @xmath100", "this surface free energy has two minima at @xmath101 and @xmath102 in the interval @xmath103 $ ] , and as such is bistable , while for @xmath104 , only the minimum @xmath101 exists , i.e. , the surface is monostable . to study the stability limit of the h phase", "we expand @xmath105 in eq .", "( [ eq13 ] ) around @xmath101 up to sixth order : @xmath106 the h phase corresponds to a local minimum of @xmath105 in eq .", "( [ eq13 ] ) if @xmath107 , where @xmath108 a standard bifurcation analysis reveals that the transition from the h phase to the han phase can be either first order or continuous .", "the transition is continuous if @xmath109 , first order if @xmath110 , and @xmath111 corresponds to a tricritical point . in the case of a first order phase transition", "the phase boundary of thermal equilibrium is given by @xmath112 we emphasize that the order of the phase transition depends only on the surface free energy @xmath113 close to @xmath101 for @xmath98 .", "therefore it is possible to have a first order phase transition even with a monostable surface characterized by a monotonic surface free energy such as the one shown in fig .", "[ fig2 ]  ( a ) for @xmath114 $ ] or the empirical equation ( [ eq14 ] ) with @xmath104 as well the more general expression @xcite @xmath115\\end{aligned}\\ ] ] with appropriate parameters @xmath116 and @xmath117 .", "first order phase transitions between the h and han texture are of particular interest for bistable liquid crystal displays . in a bistable liquid crystal display", "the two molecular configurations corresponding to light and dark states are locally stable in the thermodynamic space when the applied voltage is removed @xcite . therefore , power is needed only to switch from one stable state to another , in contrast to monostable liquid crystal displays which require power to switch and to maintain the light and the dark states .", "we now turn our attention to the case that it is not possible to evaluate @xmath118 from @xmath119 [ eq .  ( [ eq7 ] ) ] because the condition for this inversion is not satisfied [ eq .  ( [ eq9 ] ) ] . to this end", "we have chosen the parameters @xmath90 , @xmath91 , and @xmath74 for the system shown in fig .", "[ fig1 ]  ( a ) .", "figure [ fig3 ] ( a ) display ( dashed line ) and ( solid line ) while the corresponding phase diagram is shown in fig .", "[ fig3 ]  ( b ) . as is apparent from the solid line in fig .", "[ fig3 ]  ( a ) it is not possible to determine the effective surface free energy function for the all values of @xmath42 because the upper flat substrate at @xmath36 is too far away from the lower patterned substrate in order to induce all possible average anchoring orientations @xmath42 . in other words , the anchoring energy at the patterned substrate is too large to be balanced by the elastic energy for the chosen mean distance @xmath36 between the substrates ( see fig .", "[ fig1 ]  ( b ) ) . nevertheless fig .", "[ fig3 ]  ( b ) demonstrates that even this partial information about the effective surface free energy function can be used to calculate the phase diagram for cell widths sufficiently larger than the width at the critical point @xmath120 .      in the last subsection we have shown that with a suitable chemical and geometrical surface morphology on one of the interior surfaces of a liquid crystal cell", ", two stable nematic director configurations can be supported .", "the zenithally bistable nematic devices that have been studied recently @xcite consist of a nematic liquid crystal confined between a chemically homogeneous grating surface ( see fig .", "[ fig1 ]  ( c ) ) and a flat substrate with strong homeotropic anchoring .", "the profile of the asymmetric surface grating is given by @xmath121 where @xmath34 is the groove depth , @xmath66 the period , and @xmath122 is the `` blazing '' parameter describing the asymmetry of the surface profile", ". such a grating surface has been studied by brown _", "@xcite who found a first order transition between the han state , characterized by a low pretilt angle ( @xmath123 ) , and the h state , characterized by a high pretilt angle ( @xmath124 ) . strictly speaking", ", the h state does not correspond to the homeotropic texture ( see fig .", "[ fig5 ]  ( b ) below ) , but we keep the same notation as in the previous section for consistency . here", "we study phase transitions of a nematic liquid crystal in contact with the blazed surface in a more detail using the effective free energy method discussed in sec .", "ii  c.    in fig .", "[ fig : blazed ]  ( a ) the minimized free energy @xmath70 ( dashed line ) and the calculated effective surface free energy @xmath71 ( solid line ) are shown for the anchoring strength @xmath125 on the grating surface , the groove depth @xmath126 , and the blazing parameter @xmath127 .", "the effective surface free energy is asymmetric with respect to @xmath128 because of the asymmetry of the grating surface . as a consequence", "also the phase diagram , plotted as a function of the anchoring angle on the upper surface @xmath40 and the distance @xmath129 is asymmetric ( fig .  [", "fig : blazed ]  ( b ) ) .", "however , the topology of the phase diagram is the same as in the case of a symmetric substrate ( see figs .", "[ fig2 ] ( b ) , [ fig3 ] ( b ) , and for a more general discussion sec .", "iii  c below ) .", "( c ) and eq .", "( [ eq : blazed ] ) ) .", "the locally homeotropic anchoring strength on the blazed surface is @xmath130 , the groove depth is @xmath131 and the blazing parameter is @xmath127 .", "@xmath28 is the extension of the cell in the invariant @xmath26 direction and @xmath85 is the isotropic elastic constant .", "the effective surface free energy ( as well as ) is periodic with the period @xmath132 but it is asymmetric with respect to @xmath133 .", "( b ) phase diagram of the same system as a function of the anchoring angle at the upper flat substrate @xmath40 and the cell width @xmath35 with the same line code as in figs .", "[ fig2 ] and [ fig3 ] .", "the triple point ( where the han@xmath78 , han@xmath79 and h phases coexist ) is at @xmath134 and @xmath135 , and the critical points ( solid circle ) are at @xmath136 , @xmath137 and @xmath138 , @xmath139", ". for clearness , only the limits of metastability of the han@xmath79 ( 1 ) phase and the h ( 2 ) phase ( for negative @xmath40 ) are shown.,width=302 ]    ( c ) and eq .", "( [ eq : blazed ] ) ) with local homeotropic anchoring and a flat surface with strong homeotropic anchoring ( @xmath98 ) as a function of the groove depth @xmath140 and the cell width @xmath129 .", "the anchoring strength on the blazed surface is @xmath130 .", "the lines correspond to different values of @xmath122 and denote first order transitions between homeotropic ( h ) and hybrid aligned ( han@xmath79 ) phases . for small @xmath140 , the lines extend to @xmath141 corresponding to a first order _ anchoring _ transition between planar and homeotropic phases . upon increasing", "@xmath140 the first order transition lines end at critical points which are not shown in the figure .", "( b ) the energy barrier at the first order transitions with the same line code as in ( a ) .", "the lines have been obtained from the total effective free energy ( see eq .", "( [ eq6 ] ) ) , while the diamonds correspond to the energy barriers between the planar and homeotropic effective anchoring which follow from considering only the surface contribution @xmath44.,width=302 ]    ) ) for the hybrid aligned ( han@xmath79 ) phase in ( a ) and the homeotropic ( h ) phase in ( b ) on the lines of the first order transitions ( see fig .  [ fig4 ] ) for a nematic liquid crystal confined between the blazed surface ( see fig .  [ fig1 ]  ( c ) and eq .", "( [ eq : blazed ] ) ) with local homeotropic anchoring and a flat surface with strong homeotropic anchoring ( @xmath98 ) .", "the diamonds in ( a ) and ( b ) correspond to the two minima of the effective surface energy function @xmath142 .", "the model parameters and the line code are the same as in fig .", ", width=302 ]    we now concentrate on the most interesting ( from a practical point of view ) case of strong homeotropic anchoring ( @xmath86 ) at the upper homogeneous surface .", "figure [ fig4 ] ( a ) displays the phase diagram for a few values of the blazing parameter @xmath122 and a fixed value of the local homeotropic anchoring strength on the grating surface @xmath130 . for a fixed cell width @xmath129 , asymmetry ( @xmath143 ) leads to a decrease of the groove depth @xmath140 at which there is a first order transition between the han and the h phases , as compared to the nematic liquid crystal cell with the symmetric surface grating ( @xmath144 ) . upon increasing the groove depth", "@xmath140 the transition line ends at a critical point ( not shown in the figure ) , while it diverges as @xmath145 .", "the groove depth @xmath146 corresponds to an _ anchoring _ ( or surface ) transition between low tilt and high tilt _ surface _", "states which are the homeotropic and planar effective anchoring states , respectively , in the case @xmath144 .", "the effective free energy method allows one to calculate an energy barrier @xmath147 between the two bistable states , which is _ not _", "feasible by the direct numerical minimization of the free energy functional .", "the results of our calculations are shown in fig .", "[ fig4 ]  ( b ) .", "the asymmetry of the surface grating leads to a decrease of the energy barrier . with increasing groove depth @xmath140 the energy barrier decreases and eventually vanishes upon approaching the critical point .", "the diamonds in fig .", "[ fig4 ]  ( b ) denote the energy barriers for the abovementioned anchoring transitions of a nematic liquid crystal in contact with a single grating surface .", "the energy barrier between two bistable states is an important quantity for the design of a zenithally bistable nematic device .", "too small energy barrier , as compared to @xmath148 , would cause spontaneous switching between the two states because of thermal fluctuations , while enlarging the energy barrier leads to an increase of the power consumption . using the calculated values of @xmath147 ( see fig .", "[ fig4 ]  ( b ) ) one can estimate the energy barrier in a real nematic liquid crystal cell .", "for instance , for a cell of area @xmath149 and of width @xmath150 , and taking the typical values @xmath151 and @xmath152 , one obtains @xmath153 for @xmath154 and @xmath155 , which seems to be an acceptable value .", "another important quantity in zenithally bistable nematic devices is the average director orientation at the grating surface in the two degenerate states .", "the average surface director in the han ( @xmath156 ) and h ( @xmath157 ) states is shown in fig .", "[ fig5 ] for the same model parameters as in fig .", "[ fig4 ] and for the values of @xmath140 and @xmath129 corresponding to the coexistence line .", "the asymmetry of the surface grating leads to a decrease of @xmath156 and an increase of @xmath158 . for a fixed value of @xmath122", ", the difference between the two angles decreases with increasing the groove depth @xmath140 and finally vanishes upon approaching the critical point ( not shown in the figure ) .", "hence the asymmetry of the surface grating leads to a decrease of the groove depth at which the bistability is observed , which improves optical properties @xcite , and to a decrease of the energy barrier , which lowers the power consumption of a zenithally bistable nematic device . on the other hand ,", "also the difference between the two bistable states decreases which impairs optical properties of such a device .", "in section [ iiia ] we have discussed phase diagrams in the @xmath159 plane which can be described in terms of the surface free energy given by eq .  .", "however , it is instructive to consider a more general situation that the surface free energy follows from a truncation of the fourier expansion given in eq .  .", "to be able to study both symmetric and asymmetric surfaces we assume a natural generalization of eq .", ", namely @xmath160\\,,\\end{aligned}\\ ] ] which reduces to eq .   in the case of a symmetric surface characterized by @xmath161 .", "the angle @xmath162 that minimizes @xmath163 ( see eq .  ) is a function of @xmath40 and @xmath35 .", "the derivative @xmath164 is the susceptibility of the system that diverges at the critical thickness @xmath165 and remains finite and positive for @xmath107 . from eq .   the conditions for the critical angle @xmath166 follow as @xmath167 implying that @xmath120 and @xmath166 depend only on the form of @xmath113 .", "plane for the surface free energy given by eq . .", "@xmath168 and @xmath169 denote different ( usually non - uniform ) textures , the thick lines correspond to first order transitions , and black circles mark critical points .", "when @xmath11 ( defined on the unit circle @xmath170 ) has one minimum and one maximum the phase diagram can be either of type ( a ) or ( b ) .", "when @xmath11 has two minima and two maxima the phase diagram is of type ( b ) unless the minima are of equal depth , in which case it is of type ( c ) .", "note that the lines @xmath171 are identified with each other and the phase diagram can be considered as being on a cylindrical surface.,scaledwidth=49.0% ]    the extremes of @xmath11 given by eq .", "can be found easily only in the case of symmetric or antisymmetric ( @xmath172 ) surface and the same concerns the position of critical point . in this work , however , we are interested rather in possible topologies of the phase diagram in the @xmath159 plane , which result from eqs . and , and not in the exact location of critical points or transition lines . to draw schematic phase diagrams we consider @xmath11 as a function defined on the unit circle @xmath170", "depending on the parameters @xmath116 and @xmath117 , @xmath113 has either one minimum and one maximum or two minima and two maxima .", "this conclusion applies also to the function @xmath173 , thus , there can be either one or two critical points in the phase diagram ( see eqs . and ) . in the limit of large @xmath35", ", there is always a first order phase transition between two non - uniform textures corresponding to the opposite orientations of the director at @xmath21 .", "this is because @xmath174 is not a periodic function of @xmath175 at fixed @xmath40 . to find @xmath40 at the transition we expand @xmath11 around its deepest minimum ( denoted @xmath176 ) , which leads to the approximate free energy : @xmath177 where @xmath178 is the extrapolation length .", "since @xmath176 and @xmath179 are equivalent minima of @xmath11 , and both @xmath176 and @xmath40 are allowed to vary in the interval @xmath180 $ ] , the transition occurs at @xmath181 if @xmath182 or at @xmath183 if @xmath184 . with the above information we can now draw schematically the phase diagram ( see fig .  [ fig : diagram ] ) .", "since the vertical lines at @xmath171 are to be identified with each other the phase diagram can be considered on a cylindrical surface .", "away from the transition lines there is a smooth evolution from one texture to another .", "this means that at fixed @xmath35 it is possible to transform smoothly the @xmath185 texture into the @xmath186 texture , i.e. , without crossing the transition line , even for @xmath187 .", "we note that in some range of parameters , the phase diagram for @xmath11 with one minimum is topologically indistinguishable from that for @xmath11 with two minima ( fig .", "[ fig : diagram ]  ( b ) ) ; in both cases there are two critical points and a triple point .", "if @xmath11 has two equal minima ( e.g. , at @xmath101 and @xmath188 in the case of symmetric surface ) the triple point disappears and the first order transition lines extend to @xmath189 , as shown in fig .", "[ fig : diagram ]  ( c ) ."], ["we have studied the phase behavior of a nematic liquid crystal confined between a flat and a patterned substrate ( fig .", "[ fig1 ] ) using the frank - oseen model [ eq .", "( [ eq4 ] ) ] and the rapini - papoular surface free energy [ eq .", "( [ eq5 ] ) ] . an expression for the effective free energy function of the system [ eq .", "( [ eq6 ] ) ] was derived by determining an effective surface free energy characterizing the anchoring energy at the patterned surface [ eq .  ( [ eq8 ] ) ] . using the effective free energy function ,", "we have determined the phase behavior of the nematic liquid crystal confined between a flat surface with strong anchoring and a chemically patterned sinusoidal surface ( fig .", "[ fig1 ] ( a ) ) , finding first order transitions between a homeotropic texture ( h ) and hybrid aligned nematic ( han ) textures ( figs .", "[ fig2 ] ( b ) and [ fig3 ] ( b ) ) .", "it is possible to have a first order phase transition even with a monostable surface characterized by a monotonic surface free energy function ( fig .", "[ fig2 ]  ( a ) , @xmath114 $ ] ) .", "in addition we have performed direct minimizations of the original free energy functional [ eqs .", "( [ eq4 ] ) and ( [ eq5 ] ) ] on a two - dimensional grid and found remarkably good agreement with the phase boundaries resulting from the effective energy function analysis ( figs", ".  [ fig2 ] ( b ) and [ fig3 ] ( b ) ) .", "hence quantitatively reliable predictions of the phase behavior can be achieved using the effective free energy method .    using this method ,", "we have also studied the phase behavior ( fig .", "[ fig : blazed ] ( b ) ) of a nematic liquid crystal confined between a chemically uniform , asymmetrically grooved substrate ( fig .", "[ fig1 ]  ( c ) and eq .", "( [ eq : blazed ] ) ) with locally homeotropic anchoring and a flat substrate with strong homeotropic anchoring , which is a typical setup for a zenithally bistable nematic device @xcite .", "the asymmetry of the grating substrate leads to a decrease of the groove depth at which a first order transition between the h and han phases occurs ( fig .", "[ fig4 ]  ( a ) ) .", "moreover , we have determined the energy barrier between the two coexisting states ( fig .", "[ fig4 ]  ( b ) ) .", "our calculations show that the energy barrier decreases with increasing the asymmetry of the grating surface but it is well above @xmath190 for a typical nematic liquid crystal cell .", "in addition , the average director orientation at the grating surface in two bistable states has been calculated ( fig .", "[ fig5 ] ) .", "the difference between the two bistable states vanishes with increasing substrate asymmetry , which has a negative effect on the optical properties of a zenithally bistable nematic device .", "we have also generalized the model of the effective surface free energy considered by parry - jones _", "@xcite to the case of asymmetric structured substrates and obtained three possible types of the phase diagram in the plane spanned by the orientation of the director at the homogeneous surface and the thickness of the nematic cell .", "the asymmetry of the substrate causes only a shift of transition lines and critical points , compared to the symmetric case , but does not change the topology of the phase diagram .", "finally , we have verified that this model allows one to reproduce qualitatively the phase diagram of a nematic liquid crystal confined between a homogeneous planar substrate and an asymmetrically grooved surface ( fig .", "[ fig : blazed ]  ( b ) and fig .", "[ fig : diagram ]  ( b ) ) ."]]}
{"article_id": "0709.1158", "article_text": ["galaxy clusters , as the largest peaks in the cosmic density field , play an important role in astrophysics and cosmology ( e.g. * ? ? ?", "* ; * ? ? ?", "structure formation theory , realized in large scale n - body simulations , makes robust predictions for cluster space density and clustering within various cosmological models . since clusters are also the most observationally accessible features of large scale structure , they provide an opportunity to place strong constraints on both cosmological parameters and the growth of structure .", "the great challenge of cluster cosmology lies in confidently relating the dark matter halos we can robustly predict to the baryonic structures we observe .", "substantial work is being done to close the gap between theory and observations from both sides .", "on the theory side , numerical simulations of ever increasing complexity and resolution provide new insights into the evolution of baryons within clusters , and to elucidate the connection between cluster galaxies and dark matter substructure @xcite .", "observationally , much work is being done to assemble large samples of clusters detected and observed in a wide variety of ways .", "this is possible because galaxy clusters provide a rich suite of observables .", "optical light is emitted by individual cluster galaxies as well as intracluster stars .", "x - rays are emitted by both the hot intracluster medium ( icm ) and agn within cluster galaxies .", "this same hot icm scatters microwave background photons passing through the cluster , distorting their spectrum .", "finally , the total projected mass distribution of the cluster produces weak and sometimes strong lensing distortions in the images of background galaxies .", "each of these observables presents an opportunity to detect clusters and measure their properties . when combined , they allow us to cross - check our understanding of cluster physics in a variety of ways .", "clusters were first detected as early as the 18th century as anomalous groupings of similarly bright galaxies @xcite .", "optical surveys , which are relatively inexpensive , have long provided the largest cluster catalogs , primarily because of their ability to detect objects with relatively low mass thresholds . while successful in identifying thousands of clusters @xcite and providing the first evidence for dark matter @xcite , early optical detection was plagued by projection of galaxies along the line of sight @xcite .", "precise ccd photometry has enabled searches for galaxies clustered in space , brightness , and color  ( for a review of optical selection , see * ? ? ? * ) , substantially reducing the problems of projection .", "these more recent optical surveys also naturally provide accurate photometric redshifts .    with", "the advent of x - ray satellites , detection of thermal emission from the hot icm became possible @xcite .", "because x - ray emission depends on the square of the density , this provides a higher contrast that is largely immune to projection effects .", "but when spatial resolution is low , contamination from non - thermal x - ray sources can be a difficult problem", ". these can be point sources like agn , the non - thermal emission from a cooling core , or ongoing merger activity that has thrown a cluster out of thermal equilibrium . over the last several decades", ", x - ray surveys have been used to assemble a large number of cluster catalogs @xcite .", "much of our understanding of cluster physics today derives from these x - ray selected catalogs .", "substantial effort has gone into comparing optically and x - ray selected catalogs @xcite , and to comparing optical and x - ray properties of clusters to their weak lensing and sz signals @xcite .", "comprehensive comparisons of optical and x - ray properties of clusters have been hampered by the lack of large samples that are uniformly observed in both passbands .    in this paper", "we describe measurements of the x - ray properties of the largest publicly available optically - selected cluster sample : the maxbcg catalog @xcite .", "this approximately volume - limited cluster catalog spans the redshift range from @xmath2 .", "uniform optical photometry and relatively precise photometric redshifts ( @xmath11 ) for all these clusters are available from the same sloan digital sky survey  ( sdss : * ? ? ?", "* ) data from which the clusters were selected .", "x - ray observations of all these maxbcg clusters are available from the _ rosat _ all - sky survey  ( rass : * ? ? ?", "while rass exposures are too shallow to allow significant individual detections of every maxbcg cluster , they provide precise measurements of the mean x - ray luminosity ( @xmath0 ) as a function of optical richness and redshift .", "in addition , the low signal - to - noise measurements of x - ray emission from individual clusters can be used both to confirm the measurement of mean x - ray emission obtained by stacking and to provide estimates of the scatter in the optical richness  x - ray luminosity relation .", "the maxbcg catalog has been studied in a variety of complementary ways .", "for example , both dynamical  ( * ? ? ? * henceforth b07 ) and weak lensing  ( * ? ? ?", "* ; * ? ? ?", "* s07 , j07 ) observations of these clusters have been extracted from sdss data .", "these earlier observations can be combined with the mean @xmath0 measurements presented here to provide further insight into cluster physics .", "the @xmath0@xmath12 relation inferred for maxbcg clusters is described here , while the corresponding @xmath0@xmath13 relation obtained from weak lensing is discussed in a companion letter  @xcite .", "an analysis similar to that reported in this paper was performed for the nir selected 2mass cluster catalog by ( * ? ? ?", "* henceforth dkm07 ) .", "the 2mass nir flux - limited catalog has @xmath14 nearby ( @xmath15 ) groups and clusters ranging in mass from @xmath16 , selected with a matched - filter algorithm  @xcite .", "dkm07 are the first to take a large optically - selected cluster catalog and measure the mean ( stacked ) x - ray properties using rass , rather than simply cross - correlating optically - selected and x - ray - selected clusters .", "they find that the x - ray luminosity of 2mass clusters scales with optical richness ( their @xmath17 , the number of galaxies brighter than @xmath18 within @xmath19 ) .", "in addition , they derive x - ray temperatures and hydrostatic masses for the stacked 2mass clusters .    in section [ sec : input data ]", ", we briefly review the sdss data , maxbcg catalog , and rass data on which this study is based .", "section [ sec : x - ray analysis ] describes our measurement of x - ray luminosities for individual clusters , as well as our methods for determining the mean x - ray luminosity of a set of clusters with similar richness .", "we describe the mean relation of x - ray luminosity as a function of richness , scatter in this relation , and the underlying median @xmath20 relation in section [ sec : meanrelation ] .", "section [ sec : biases ] discusses several possible sources of systematic bias in these results . in section [ sec : lsig ] we combine these results with dynamical measurements of maxbcg clusters to produce a measurement of the @xmath21 relation .", "conclusions and some discussion of future work are presented in section [ sec : summary ] . throughout this work we use a @xmath22cdm cosmology with @xmath23 and @xmath24 .", "the measurements described here are based on two wide - area imaging surveys ; sdss and rass .", "galaxy clusters are selected from the sdss five - band imaging data using a red sequence selection method .", "sdss data also allow measurement of cluster redshifts and richnesses .", "x - ray emission from these clusters is then measured by from the rass photon maps . in this section", "we describe briefly the sdss imaging data , galaxy cluster selection and calibration , and rass input data .", "optical data for this study are drawn from sloan digital sky survey : a combined imaging and spectroscopic survey of 10@xmath25 deg@xmath26 in the north galactic cap and a smaller region in the south .", "the imaging survey was carried out in drift - scan mode in five sdss filters ( @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 ) to a limiting magnitude of @xmath32 @xcite .", "photometric errors are typically limited at bright magnitudes by systematic uncertainties at the 3% level .", "the spectroscopic survey targets both a `` main '' sample of galaxies with @xmath33 and a median redshift of @xmath34 @xcite and a `` luminous red galaxy '' sample @xcite which is approximately volume limited out to z=0.38 . for more details of the sdss", "see @xcite and @xcite .", "the maxbcg cluster catalog is selected from imaging data contained in dr5 of the sdss .", "selection of galaxy clusters from this imaging data is done using the `` maxbcg '' algorithm .", "details of the algorithm are presented in @xcite , while the catalog and a description of its properties may be found in @xcite . in brief , the algorithm exploits two well - known features of rich galaxy clusters in addition to the tight spatial clustering of cluster galaxies .", "first , the bright end of the cluster luminosity function is dominated by red sequence galaxies with a small dispersion in color - magnitude space ( the e / s0 ridgeline ) .", "second , clusters contain a distinct brightest cluster galaxy ( bcg ) located near the center of the galaxy distribution .", "while some clusters lack an obvious central , dominant galaxy , every cluster does possess some red sequence galaxy brighter than any other .", "the algorithm measures independently the likelihood that a galaxy is spatially located in an overdensity of e / s0 ridgeline galaxies with similar @xmath35 and @xmath36 colors , and that it has the color and magnitude properties of a typical bcg .", "both likelihoods are evaluated for every sdss galaxy at a grid of redshifts .", "the redshift which maximizes the product of these likelihoods is then found for each galaxy .", "for e / s0 galaxies , this corresponding maximum likelihood redshift then provides a good estimate of the cluster redshift .", "once this list of cluster center likelihoods is assembled , these potential centers are ranked by decreasing maximum likelihood . the first cluster is seeded on the highest likelihood center , that cluster s bcg .", "galaxies projected within a scaled radius , @xmath37 , of this bcg and within @xmath38 in @xmath31 are eliminated from the list of potential centers .", "@xmath37 is the radius interior to which the mean density is 200 times the critical density ( @xmath39 ) as determined from sdss galaxy populations  @xcite .", "the process is repeated for the next most likely bcg on the list , given that it has not been eliminated by a higher likelihood bcg , and continues likewise until all potential centers have either been labeled as cluster bcgs or have been subsumed by higher likelihood centers .", "each cluster defined in this way has a center defined as the bcg location , an estimated redshift , and a richness , @xmath6 , given by the number of e / s0 ridgeline members falling within @xmath37 of the bcg and brighter than 0.4 @xmath40 .", "the final cluster catalog contains an array of measured properties , including photometric redshifts , richnesses , optical luminosities , and locations .", "the public maxbcg catalog contains a total of 13,823 clusters drawn from approximately 7500 square degrees of sky between redshifts of 0.1 and 0.3 , with a median redshift @xmath41 .", "the center for each cluster is defined as the location of the bcg identified by the algorithm .", "the richness of the cluster , @xmath6 , ranges between 10 and 188 in the public catalog , and in principle extends down to @xmath42 where the maxbcg selection function is less well - understood . for this study", "we include some clusters of lower richness , adding an additional 3532 clusters with @xmath43 .", "this slightly extended catalog allows us to use the same richness bins studied in the analysis of maxbcg galaxy dynamics  ( b07 ) and gravitational lensing  ( s07 , j07 ) .", "redshift estimates for the clusters , produced as part of the cluster finding process , have been shown by comparison to spectroscopic redshifts to be quite accurate , with @xmath11 @xcite .", "catalog completeness and purity have been studied in some detail in @xcite ; both are quite high .", "completeness is estimated to be @xmath44 for clusters with masses greater than @xmath45 , and purity is @xmath44 for clusters with richnesses @xmath46 .    the cluster population in this catalog has been used to derive constraints on cosmology @xcite using cluster counts .", "the relationship between the maxbcg richness @xmath6 and mass has been studied through galaxy dynamics  ( b07 ) and weak lensing ( s07 , j07 )", ". further work on galaxy populations , mass - to - light ratios , and improved richness estimates for these clusters is in progress .", "the _ rosat _ all - sky survey  ( rass , * ? ? ?", "* ) took place primarily during a six - month campaign in 1990 - 1991 to image the whole sky in soft x - rays ( 0.1 - 2.4 kev ) with the rosat position sensitive proportional counter  ( pspc , * ? ? ?", "the survey scanned the sky in great circles , with the largest net exposure time ( @xmath47 ) near the ecliptic poles .", "the typical field coincident with the maxbcg survey region , which does not overlap the northern ecliptic pole , has an effective exposure time of @xmath48 .", "the point spread function ( psf ) for these rass scans is very broad ( full - width - half - maximum of @xmath49 ) and is dominated by far off - axis photons due to the survey strategy .", "@xcite released reprocessed photon maps and exposure maps of the entire rass survey region .", "these photon maps provide the input for the analysis described in this paper .", "rass data has been used to create several catalogs of purely x - ray selected objects .", "the rosat bright source catalog  ( bsc , * ? ? ? * ) consists of 18,811 sources with a typical signal - to - noise @xmath50 .", "the position resolution is superior to the psf fwhm with 68% ( 90% ) of the sources within @xmath51 ( @xmath52 ) .", "in addition to the bsc , there is a companion faint source catalog  ( fsc , * ? ? ? * ) consisting of 105,924 sources with a typical signal - to - noise @xmath53 .", "most are unresolved .", "combined , these catalogs provide soft x - ray detected sources which can also be compared to the maxbcg clusters .    the rass photon data has also been used as an input for x - ray flux - limited cluster catalogs . the brightest cluster sample  ( bcs , * ? ? ?", "* ) is a flux - limited sample of the brightest 201 clusters in the northern hemisphere with fluxes @xmath54 .", "the northern rosat all - sky galaxy cluster survey  ( noras , * ? ? ?", "* ) is a catalog of 378 extended x - ray sources that have been confirmed to be clusters via optical follow - up . due to the broad rosat psf", ", this catalog is only @xmath55 complete at their stated flux limit , @xmath56 , which corresponds to a luminosity @xmath57 at our median redshift @xmath41 . @xcite", "have performed an initial comparison between noras clusters and maxbcg clusters and have found that the maxbcg detects @xmath58 of noras objects .", "this is consistent with estimates of the completeness from simulations  @xcite .", "the _ rosat_-eso flux limited x - ray galaxy cluster survey  ( reflex : * ? ? ?", "* ) is similar to noras in the southern sky , with the same flux limit .", "this catalog of 447 clusters is over 90% complete due to improvements in rass analysis , although only a small fraction of reflex overlaps the maxbcg survey area .", "the typical rass exposure time for maxbcg clusters , 400 s , is too short to allow significant detections for individual clusters .", "the large number of maxbcg clusters , however , allows us to make up for this .", "for example , there are 7986 clusters with richnesses 9 @xmath59 11 . for these objects ,", "the total rass exposure time is @xmath60 .", "such a large total exposure allows us to measure the mean x - ray emission from these clusters quite precisely .", "details of the stacking method are described later in this section , and outlined here .", "we begin by dividing the clusters into nine richness ( @xmath6 ) bins . to simplify comparison to other maxbcg analyses", ", we use the same richness bins used in measurements of the mean velocity dispersions  ( b07 ) .", "the number of clusters in each richness bin is shown in table  [ tab : ncluster ] .", "cccccc @xmath61 & 92.85 & 0.21 & 1727 & 55 & 55 + @xmath62 & 58.22 & 0.21 & 1469 & 146 & 140 + @xmath63 & 44.67 & 0.21 & 1317 & 207 & 201 + @xmath64 & 35.74 & 0.21 & 1201 & 356 & 339 + @xmath65 & 28.57 & 0.22 & 1102 & 665 & 633 + @xmath66 & 22.70 & 0.22 & 997 & 1128 & 1060 + @xmath67 & 18.91 & 0.23 & 941 & 1141 & 1099 + @xmath68 & 13.88 & 0.23 & 823 & 5651 &", "5405 + @xmath69 & 9.80 & 0.23 & 727 & 7986 & 7566 + random & & & 7986 & 7529 +    as we do not have x - ray centers for individual clusters , we treat the bcg selected by the maxbcg algorithm as the center of each cluster and stack on these centers .", "in @xcite it was shown that most maxbcg - selected bcgs ( @xmath70 ) agree well with the center of x - ray selected clusters , which has also been seen for other optically - selected catalogs  ( e.g. * ? ? ?", "possible biases introduced by this assumption are discussed in section  [ sec : centering ] .", "every source and background photon in the stacked analysis is scaled and weighted to the median redshift of the clusters in the catalog , @xmath71 . with these weighted photon maps", "we construct stacked images in section  [ sec : images ] , radial profiles in section  [ sec : profiles ] , and spectra and luminosities in section  [ sec : spectra ] .", "we obtain rass photon data and merged exposure maps from the archives available at the high energy astrophysics science archive research center ( heasarc ) .", "the rass data and exposure maps are distributed in arbitrarily constructed @xmath72 degree fields with significant overlap . nevertheless , since we utilize a large background annulus extending to @xmath73 from the central bcg , a significant fraction ( @xmath74 ) of the maxbcg clusters fall across a field boundary .", "we have therefore built a tool that extracts all rass photons in a given aperture , from multiple rass fields if necessary , counting only once those photons which appear multiple times in overlap regions .", "after photon retrieval , the appropriate merged exposure maps ( mex file ) are used to calculate the effective exposure time ( @xmath75 ) at the position of each detected photon .    before stacking", "we exclude from the list a subset of maxbcg clusters which might bias our x - ray measurements for unphysical reasons .", "no cut on galactic absorption is required , as the sdss observations are restricted to high galactic latitude .", "the typical equivalent galactic hydrogen column density is a very low : @xmath76  @xcite for the maxbcg clusters , with a maximum of @xmath77 .", "although the typical rass exposure time is @xmath78 , there are some fields with significantly less .", "this makes source and background estimation difficult , while not adding significantly to the signal . for this reason", ", we reject all clusters with less than 200  s mean exposure time .", "this removes 4% of the total number of clusters , but less than 1% of the net exposure time .", "an important possible source of contamination for stacked cluster measurements is the presence of a few extremely bright foreground sources .", "we would like to reject regions contaminated by these sources .", "rass images of maxbcg clusters , especially at higher redshift , are often unresolved , making it difficult to remove point sources using extent information , so we proceed as follows .", "we first identify those rosat bright source catalog ( bsc ) sources in the survey area with count rates higher than that expected for emission from any cluster at redshift beyond 0.1 .", "this corresponds to a count rate of @xmath79 .", "there are 179 bsc sources in the maxbcg survey area with rosat soft - band ( 0.1 - 0.5 kev ) or hard - band ( 0.5 - 2.1 kev ) count rates above this limit . of these 179 sources ,", "only 13 are clusters as identified in the _ rosat _ brightest cluster sample  ( bcs , * ? ? ?", "* ) . of these 13 clusters ,", "only one of these objects is associated with a maxbcg cluster ( abell 2142 , the richest and one of the nearest clusters in the catalog ) .", "the other high - flux clusters are at redshifts @xmath80 .", "further visual inspection has confirmed that abell 2142 is the only maxbcg cluster associated with one of these extremely bright sources . for the stacking analysis", ", we reject all maxbcg clusters ( except for abell 2142 ) within @xmath81 of any of these bright sources , due to possible contamination from non - cluster photons .", "this removes only 0.6% of the maxbcg catalog .", "the virgo cluster is another important foreground source .", "this nearby , x - ray bright cluster dominates the rass emission over a significant region of sky , thus making it difficult to estimate the x - ray emission from background clusters .", "we therefore reject all clusters that fall within @xmath82 degrees of m86 or m87 .", "this is a relatively small effect , removing only 0.1% of the maxbcg clusters .", "table  [ tab : ncluster ] shows the number of clusters in each richness bin before and after the rejections outlined in this section , as well as the mean richness for each bin .    to stack the x - ray measurements of the clusters , rass photons from each cluster", "must be scaled and appropriately weighted . the projected physical distance to the bcg", "is calculated for each photon , and this scaled distance is used in image construction and radial profile calculations . each source and background photon", "is then weighted by a factor of @xmath83 , where @xmath84 is the luminosity distance to the cluster and @xmath85 is the luminosity distance to the median cluster redshift @xmath71 . for background estimation", "we use a fixed annulus with inner ( outer ) radius of @xmath86 ( @xmath87 ) , similar to that used for the noras catalog for individual clusters  @xcite .", "this background annulus corresponds to a physical distance of @xmath88 at our minimum redshift of @xmath89 and a @xmath90 at our maximum redshift of @xmath91 .", "we confirm that our results are not sensitive to the exact choice of background annulus . to calculate the weighted count rate ( @xmath92 ) in a given bin ( image , radial , or spectral ) for a cluster at redshift @xmath31 , we follow a procedure similar to @xcite , where @xmath93 , where @xmath75 is the exposure time at the position of each detected photon .    as a simple test of our stacking procedure", ", we also extract rass photons from a set of positions selected at random from the region covered by the maxbcg catalog .", "each such random position is associated with a real cluster redshift drawn from the catalog in the @xmath69 bin .", "the random points are analyzed identically to the real cluster positions .", "the background - subtracted stacked images are shown in figure  [ fig : imageplots ] .", "each image has a projected radius of @xmath94 .", "the images contain photons from the _ rosat _ hard band ( channels 51 - 201 , 0.5 - 2.0 kev ) as this band has the highest signal - to - noise for cluster emission .", "after background subtraction , the counts were put in @xmath95 bins . for display , each image has been scaled with histogram equalization to show similar background noise levels , as there are two orders of magnitude more clusters in the poorest bin than the richest bin .", "the contours are drawn @xmath96 and @xmath97 above background level .", "highly significant x - ray emission is seen in each stack of clusters .", "by contrast , in a stack of 7,566 random points there is no significant excess , and no sign of emission centered in the stack .", "it is worth noting that the average count rate in the @xmath69 bin is only @xmath98 , corresponding to a flux @xmath99 .", "there are a total of @xmath100 in excess of background from @xmath101 clusters ; an average of only 1 source count per cluster .", "this illustrates the power of the stacking exercise to probe very low count rates , and allows us to use rass to measure very low @xmath5 cluster emission even at a moderate redshift of @xmath102 .", "figure  [ fig : profiles ] presents the background - subtracted surface - brightness profiles for the stacked x - ray images in each richness bin .", "the radial profiles have been summed in @xmath103 ( @xmath104 ) bins inside ( outside ) @xmath105 . for reference", ", the dotted lines represent the @xmath106 background level .", "significant cluster emission is seen in each richness bin . for the richest bins ,", "this extends out beyond @xmath105 .", "the x - ray surface brightness profiles for the different bins look remarkably similar , except for the normalization which increases strongly with richness and the signal - to - noise which decreases slowly with richness .", "we fit the surface brightness profiles out to @xmath107 with a standard @xmath108 model , @xmath109 .", "the best - fit parameters are presented in table  [ tab : betafits ] , and are overplotted with dashed lines in figure  [ fig : profiles ] .", "the @xmath108 model results in a good fit in all cases .", "cccc @xmath61 & @xmath110 & @xmath111 & 37.7/72 + @xmath62 & @xmath112 & @xmath113 & 43.2/72 + @xmath63 & @xmath114 & @xmath115 & 36.2/72 + @xmath64 & @xmath116 & @xmath111 & 75.1/72 + @xmath65 & @xmath117 & @xmath118 & 46.3/72 + @xmath66 & @xmath119 & @xmath120 & 88.1/72 + @xmath67 & @xmath121 & @xmath122 & 38.9/72 + @xmath68 & @xmath117 & @xmath118 & 36.2/72 + @xmath69 & @xmath123 & @xmath124 & 63.4/72 +    the core radii , @xmath125 , for these @xmath108 model fits are surprisingly large , ranging from @xmath126 to @xmath127 .", "these are significantly larger than those typically seen for x - ray clusters  ( @xmath128 , e.g. * ? ? ?", "the @xmath108 parameters are typical for x - ray clusters  ( e.g. * ? ? ?", "* ; * ? ? ?", "there is a slightly significant ( @xmath129 ) trend in @xmath108 with richness , such that the richer clusters tend to have slightly larger values of @xmath108 .", "this slight trend is in the same sense as has been seen in previous work correlating @xmath108 with cluster mass  @xcite .", "these @xmath108-model parameters are similar to those measured by dkm07 , in which @xmath108 was not seen to be strongly correlated with richness .", "these inconclusive results are caused by two effects ; the broad point spread function for rass objects , and the offset distribution between the bcg and the x - ray cluster emission .      to study the effect of the rass psf on the stacked profiles , we test our analysis on bright rosat point sources .", "these sources were selected from the wga catalog  @xcite , which contains 88,621 well measured point - like sources selected from rosat pspc pointed observations .", "we take @xmath130 moderately bright ( @xmath131 ) point sources that overlap the sdss dr5 mask .", "these bright sources are nearby galactic sources as well as agn and quasars at moderate and high redshift , and are also well detected in rass .", "each point source is randomly assigned a redshift drawn from the maxbcg cluster catalog redshift distribution .", "we note that a nearby unresolved point source is indistinguishable from a distant unresolved point source .", "the point sources are then stacked in exactly the manner described for clusters above .", "figure  [ fig : ptsrc ] shows the radial profile of the stacked point sources ( diamonds ) , which is essentially a measure of the rass psf with the radial scaling calculated for the maxbcg cluster redshifts . while this radial profile is much more sharply peaked than that of the stacked clusters , the broad rosat psf scatters significant x - ray emission to distances @xmath132", "( @xmath133 at @xmath134 ) from the center .", "this psf smearing implies an effective minimum on the @xmath108-model core radius , as well as creating a possible bias in the calculation of the @xmath108 parameter .", "we note that @xmath135 of the selected wga point sources are within @xmath136 of maxbcg clusters .", "excluding the point sources that are neighboring these possibly extended x - ray sources does not significantly alter this result .", "the stacked , projected surface brightness profile may also be affected by our choice of cluster centers .", "there are several ways in which the bcg chosen by the maxbcg algorithm might be offset from the x - ray emission from the cluster .", "dynamically active clusters and clusters in the process of merging can have a significant offset between the x - ray centroid and the bcg .", "there is also the possibility that the maxbcg algorithm chose an incorrect center ( this is addressed more fully in section  [ sec : centering ] ) .", "in addition , there may be additional x - ray point sources such as agn that are associated with the clusters .", "all of these effects result in an effective optical / x - ray offset distribution that may bias the radial profiles .", "we model the optical / x - ray offset distribution by matching the maxbcg catalog to known x - ray sources . for this exercise ,", "we use the fact that many maxbcg clusters ( @xmath137 ) are associated with individual detections in the rosat bright source and faint source catalogs .", "most of these bsc and fsc sources have not been previously recognized as associated with clusters , primarily because they are too faint or too distant to be seen as significantly extended in rass .", "we match the maxbcg clusters to the bsc and fsc catalogs , allowing multiple x - ray sources to match to each cluster .", "this ensures that we find all x - ray sources associated with each cluster , as well as all possible random matches .", "we similarly match an equal number of random locations drawn from the maxbcg survey region to the bsc and fsc catalogs .", "figure  [ fig : ptmatchhist ] shows the projected offset distribution from the maxbcg clusters to the bsc and fsc catalogs .", "the solid ( dashed ) line shows a histogram of maxbcg ( random ) offsets in @xmath103 bins .", "the dotted line shows the residual distribution after subtracting the random matches from the cluster matches .", "the subtracted histogram in figure  [ fig : ptmatchhist ] shows a large excess of x - ray sources associated with the optical cluster centers .", "there is a tight core in which the bcg is within @xmath138 of an x - ray source , as well as a long tail extending out to @xmath139 .", "the x - ray source excess at large radius is likely to be associated with the maxbcg clusters .", "these sources comprise a mix of merging clusters , clusters with poorly identified centers , and associated point sources such as agn .", "the subtracted histogram is used as a first - order empirical radial distribution of x - ray sources associated with maxbcg cluster centers .", "we now estimate the additional effect on the radial profile due to the radial offset distribution of x - ray sources associated with the maxbcg clusters . we take each wga point source from section  [ sec : rasspsf ] and", "randomly alter its position so that the distribution is offsets is identical to the empirical radial distribution .", "although we have only measured the offset distribution for the brightest @xmath140 of maxbcg clusters , we assume this is a good proxy for the offset distribution between the x - ray emission and bcgs of all the maxbcg clusters .", "these offset point sources are run through our stacking analysis .", "figure  [ fig : ptsrc ] shows the radial profile of the offset point sources after stacking ( circles ) .", "this radial profile is reasonably well fit with a @xmath108 model , although the fit is poor within @xmath141 . at large radii", "the radial profile is well fit with @xmath142 .", "thus , at large radii it is not possible to distinguish between true _ extended _ cluster emission and x - ray _ point _", "sources convolved with the observed maxbcg ", "x - ray source offset distribution .", "this calls into question the utility of the @xmath108 model parameterization for this exercise .", "the stacked x - ray profile is a convolution of the rosat psf , the centering distribution , and the true extended x - ray emission .", "operationally , it is not possible to separate the contributions from these three components .", "therefore , the @xmath108 parameters that are the result of a stacking exercise such as this one , where most of the individual x - ray clusters are not detected , should be used cautiously .", "we perform spectral fitting of the photons in each stack to calculate x - ray luminosities .", "the spectra were analyzed using xspec version 11.3.2  @xcite .", "spectral analysis on the stacked rosat data involves a variety of complications , some of which are discussed in dkm07 .", "we outline some key issues here .", "all rass observations integrate exposure time across the _ rosat _ field - of - view . as a result", ", we use the spectral response file pspcc_gain1_256.rmf from heasarc , suitable for off - axis observations .", "when combined with the appropriate ancillary response file ( arf ) , we obtain the vignetting corrected spectral response for the off - axis photons that comprise our spectra .", "however , the exposure times extracted from the merged exposure maps for rass are already corrected for vignetting , using the mean spectrum of the x - ray background in the pspc band . in practice", ", this means that our spectra , fit with the standard rmf+arf combination , are corrected for vignetting _", "twice _  once by using the exposure times obtained from the merged exposure map , and once in the calculation of the arf file . by comparing the on - axis response file rsp from heasarc with the combination of the rmf+arf we can approximate the extra vignetting correction .", "this correction is a function of energy , and depends ( weakly ) on the spectrum of the observed source . by simulating cluster spectra of various temperatures ,", "we calculate the typical correction factor averaged across the rosat hard band ( 0.5 - 2.0 kev ) which is dominated by cluster flux .", "this correction factor is @xmath143 . using this correction factor results in very good agreement between our spectral analysis and the reflex count - rate to flux conversion tables of @xcite .", "we estimate that applying this correction factor to all flux and luminosity values determined from stacked spectra adds an additional systematic error of @xmath144 .    for our analysis", ", we use the rmf file described above , combined with an arf output with the ftool pcarf .", "as noted in dkm07 , each arf file is essentially identical , because each stacked cluster from rass samples photons from the entire rosat field of view .", "spectral files are then grouped with a minimum of 50 counts per bin ( after background subtraction ) to ensure valid results using @xmath145 statistical analysis .", "fits performed with xspec were restricted to the 0.1 - 2.1 kev range .", "the uncertainties in spectral fit parameters are 90% confidence errors , obtained by allowing all fit parameters to vary simultaneously .", "we obtain cluster spectra by summing all weighted cluster photons in both fixed physical apertures and scaled apertures of the optically determined @xmath37 ( see section  [ sec : maxbcg ] ) .", "the fixed aperture of @xmath146 is chosen as a reasonable fiducial value because this provides good signal - to - noise , and because the radial profiles do not appear to change significantly with richness .", "the @xmath37 value for each bin was taken as the median @xmath37 of all the clusters in the richness bin ; these values are reported in table  [ tab : lxr200 ] .", "background spectra are stacked with the same weights as the source spectra , using the annuli defined in section  [ sec : extraction ] .", "as described previously , each photon is weighted to the median redshift @xmath147 .", "unfortunately , @xmath148-correction of individual photons is not possible , due to the fact that the detection channel of the incident photons is most strongly dominated by the spectral response of the rosat instrument .", "the stacking procedure will therefore tend to `` smear out '' the incident cluster spectra .", "simple simulations of x - ray spectra using xspec show that this does not create a large effect on our best - fit spectral values , as our redshift range of @xmath149 is not particularly broad .", "similarly , we do not make any corrections for possible redshift evolution in x - ray luminosity ; we will address this further in section  [ sec : redshift ] .", "ccccccc @xmath61 & 92.85 & 1.73 & @xmath150 & @xmath151 & @xmath152 & 47.2/76 + @xmath62 & 58.22 & 1.47 & @xmath153 & @xmath154 & @xmath155 & 61.5/77 + @xmath63 & 44.67 & 1.32 & @xmath156 & @xmath157 & @xmath158 & 87.4/65 + @xmath64 & 35.74 & 1.20 & @xmath159 & @xmath160 & @xmath161 & 100.6/95 + @xmath65 & 29.57 & 1.10 & @xmath162 & @xmath163 & @xmath164 & 62.9/95 + @xmath66 &", "22.70 & 0.99 & @xmath165 & @xmath166 & @xmath167 & 113.2/97 + @xmath67 & 18.91 & 0.94 & @xmath168 & @xmath166 & @xmath169 & 77.3/88 + @xmath68 & 13.88 & 0.82 & @xmath170 & @xmath171 & @xmath172 & 182.3/134 + @xmath69 & 9.80 & 0.73 & @xmath173 & @xmath174 & @xmath175 & 127.8/112 +    ccccccc @xmath61 & 92.85 & 1.73 & @xmath176 & @xmath177 & @xmath178 & 40.5/61 + @xmath62 & 58.22 & 1.47 & @xmath179 & @xmath180 & @xmath181 & 40.1/59 + @xmath63 & 44.67 & 1.32 & @xmath182 & @xmath183 & @xmath184 & 85.6/61 + @xmath64 & 35.74 & 1.20 & @xmath185 & @xmath186 & @xmath187 & 92.0/81 + @xmath65 & 28.57 & 1.10 & @xmath188 & @xmath189 & @xmath190 & 78.1/84 + @xmath66 & 22.70 & 0.99 & @xmath191 & @xmath192 & @xmath193 & 86.0/94 + @xmath67 & 18.91 & 0.94 & @xmath194 & @xmath195 & @xmath196 & 73.6/76 + @xmath68 & 13.88 & 0.82 & @xmath197 & @xmath171 & @xmath198 & 182.3/134 + @xmath69 & 9.80 & 0.73 & @xmath199 & @xmath200 & @xmath201 & 137.6/114 +    the spectra are fit with an absorbed thermal plasma model  @xcite .", "the metallicity is fixed at 0.3 solar  ( e.g. * ? ? ?", "* ) , and the redshift is set to the median scaled redshift of 0.2296 .", "the luminosities are calculated in the rest - frame 0.1 - 2.4  kev band at the median redshift .", "we prefer to calculate the 0.1 - 2.4  kev luminosity rather than bolometric luminosity ( @xmath202 ) , due to the large extrapolations required to obtain bolometric values , which strongly depend on spectral temperatures ( see below ) .", "the best - fit spectral parameters for the scaled @xmath37 apertures are shown in table  [ tab : lxr200 ] , and the parameters for the fixed @xmath146 apertures are shown in table  [ tab : lx750 ] .    in order to account properly for the variations in cluster luminosity in each richness bin", ", we use bootstrap resampling to estimate the luminosity errors . in each richness bin , we run 2000 trials by sampling the same number of clusters in that bin with replacement . to save considerable time with processing , we did not recreate the entire stacking procedure for each resampling .", "instead , we take the individual cluster counts , scaled to the median redshift .", "we confirm that the average scaled count rate is a good proxy for luminosity in the 0.1 - 2.4 kev band ( see section  [ sec : scatter ] for details on this calculation ) .", "the 68% ( @xmath106 ) confidence interval obtained from the bootstrap is added in quadrature with the @xmath5 errors obtained from the spectral fits .", "only the richest ( @xmath61 ) bin is dominated by the bootstrap error calculation , due to the large range in richness and @xmath5 in the bin .", "figure  [ fig : ltplot ] shows the mean @xmath0@xmath203 relation for our stacked clusters ( within the scaled @xmath37 aperture ) , compared to the @xcite @xmath5@xmath204 ( 0.1 - 2.4 kev ) relation .", "we see the cluster temperature increase with luminosity , but our observed @xmath0@xmath203 relation is noticeably steeper than that of @xcite .", "the stacked x - ray temperatures appear to underestimate the expected temperature , especially at @xmath0 larger than a few times @xmath205 .", "this discrepancy highlights the challenge of measuring cluster temperatures with rosat , as well as the challenges of measuring stacked cluster temperatures .", "first , rosat has sensitivity only to soft x - rays .", "when the break of the bremsstrahlung spectrum , determined by the temperature of the hot gas , is above @xmath206 , then the x - ray temperature becomes difficult to constrain .", "the _ rosat", "_ temperatures of hot clusters are generally underestimated ; a particular example is abell 1689 .", "this bright cluster has a nearly isothermal profile in _", "xmm / newton _ observations with a temperature of @xmath207 ( consistent with the asca and _ chandra _ values ) , but has a best - fit _ rosat_/pspc temperature of @xmath208  ( e.g. * ? ? ?", "second , when we stack many non - isothermal clusters with different temperatures , our isothermal raymond - smith spectrum is no longer appropriate .", "for example , @xcite have shown for _ chandra _ observations how the spectroscopically - weighted temperature can differ significantly from the emission - weighted temperature for single clusters , with the better measured colder gas dominating the spectral fit .", "when these effects are combined , interpreting the average best - fit spectral temperature from a stack of hundreds of clusters is non - trivial .", "most importantly for our task of measuring the mean @xmath0 , the rosat luminosity in the 0.1 - 2.4 kev band is virtually insensitive to the cluster temperature .", "if we fix the spectral temperature to the value predicted by the @xcite @xmath5@xmath204 relation , then @xmath0 changes by @xmath209 in each bin , with no systematic bias with richness .", "similarly , we can follow the prescription of @xcite to convert the mean 0.5 - 2.0 kev rass count rate to flux and luminosity .", "this method assumes the @xcite @xmath5@xmath204 relation , and requires iteratively calculating the flux and temperature .", "the @xmath0 values thus obtained are consistent with those from the spectral fitting to within @xmath210 .", "all this suggests that calculating 0.1 - 2.4  kev @xmath5 obtained from _", "rosat_which relies primarily on photon counting  is more robust than calculating @xmath204 and extrapolating to @xmath202 .", "although a small @xmath148-correction ( @xmath144 ) is required to extrapolate from observer frame 0.1 - 2.4  kev to rest - frame luminosity , this correction is not very sensitive to the spectral temperature at the moderate redshift of the maxbcg clusters .", "we stack the clusters and calculate the 0.1 - 2.4 kev x - ray luminosity as described in the previous section .", "figure  [ fig : lxn200plot ] shows the mean @xmath0@xmath1 relation , both for a fixed @xmath146 aperture and for the scaled @xmath37 apertures .", "we find a strong correlation between @xmath0 and @xmath1 with both the fixed and the scaled apertures . in the low richness bins ( @xmath211 ) , the scaled aperture @xmath37 is approximately equal to the fixed @xmath146 aperture , and @xmath0 is approximately the same for both choices of aperture . in the high richness bins ,", "the scaled aperture @xmath37 is significantly larger than @xmath146 , yielding a slightly larger @xmath0 for the scaled apertures .", "for this reason , the @xmath0@xmath1 relation is slightly steeper using scaled apertures .", "the best - fit mean relationships are :    @xmath212    @xmath213    power law fits are chosen to pivot around @xmath214 to approximately decouple errors in slope and normalization .", "dkm07 also found that the mean x - ray luminosity of the 2mass clusters scales with optical richness ( @xmath17 ) . a direct comparison between the dkm07 relation and our relation is challenging , due to the different luminosity and richness definitions we employ , but a first - order conversion is possible .", "first , we convert the dkm07 bolometric luminosities ( their table 2 ) to _ rosat _ 0.1 - 2.4 kev luminosities using table 5 of @xcite . ]", "re - fitting for the @xmath0@xmath17 relation , we find @xmath215.@xmath17 relation does not deviate from a power - law at any richness , unlike the @xmath216@xmath17 relation in figure 10 of dkm07 .", "] as the 2mass catalog is not yet public , we can not make a direct comparison of @xmath6 to @xmath17 for individual clusters .", "futhermore , most of the clusters in the 2mass catalog are at @xmath217 , while the maxbcg catalog covers @xmath218 . at this time ,", "the most we can say is that the slope of the @xmath0richness relation is roughly similar between our work and dkm07 .", "comparisons to x - ray selected catalogs ( see section  [ sec : xrayselect ] ) as well as prior studies  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ) lead us to expect significant scatter in the x - ray luminosity  richness relation for individual clusters . with a large scatter ,", "the mean statistic used to calculate the stacked @xmath0-@xmath1 relation may be significantly biased relative to the median relation . to understand this bias", ", we first assume that there is a log - normal conditional probability @xmath219 for the x - ray luminosity of a cluster at richness @xmath6 : @xmath220,\\ ] ] with constant intrinsic scatter @xmath221 , and mean log luminosity that follows : @xmath222 where @xmath223 is the log normalization of @xmath5 at @xmath214 , and @xmath224 is the slope of the @xmath5@xmath6 relation .", "note that @xmath225 for a log - normal distribution is equivalent to the median ( and geometric mean ) of the distribution . for the duration of this paper", ", we employ the notation @xmath226 .    at a given richness", ", we wish to measure the median x - ray luminosity @xmath227 , the peak of the underlying log - normal distribution", ". however , the stacking exercise we have undertaken is fundamentally a calculation of the _ arithmetic mean _ ( @xmath0 ) at a given @xmath1 . for a log - normal distribution with median @xmath227 and intrinsic scatter @xmath221 ,", "the arithmetic mean is @xmath228 .", "thus , the stacked normalization is an overestimate of the median of the underlying distribution by a factor of @xmath229 .", "if the scatter is large , the stacked ( mean ) normalization will be dominated by the most luminous clusters and will be biased high .", "for example , an @xmath230 scatter indicates a @xmath231 bias .    to constrain this scatter ,", "we begin with measurements of x - ray flux ( and hence @xmath5 ) at the locations of each cluster with @xmath232 .", "though many of these are low signal - to - noise detections , they can be used to measure scatter .", "the idl astronomy library tool linmix_err is used to fit @xmath5 as a function of @xmath6 with intrinsic scatter @xmath221  @xcite .", "this tool uses a bayesian approach to linear regression with errors in x and y and is well behaved even when the measurement errors dominate .", "it also handles non - detections and upper limits in y. monte carlo simulations show that the selected @xmath6 cut provides a large enough richness range to constrain the slope and scatter , and provides larger signal - to - noise than the entire cluster catalog , as over @xmath230 of the 955 clusters with @xmath233 are detected at at least the @xmath106 level .", "furthermore , the fits are not sensitive to the precise richness cutoff chosen .", "to calculate the @xmath5 for each of these often marginally - detected clusters , we use a procedure based on the count - rate to flux conversion method from the reflex survey  @xcite .", "first , we calculate the 0.5 - 2.0 kev ( rosat hard channel ) count rate in a @xmath146 aperture ; the fixed aperture was used because it results in better signal - to - noise than a larger scaled aperture ( where we only see significant signal in the stacked profiles ) .", "the local background is calculated in a @xmath234 annulus using the sector - rejection method of @xcite .    in reflex the temperature", "is estimated in an iterative fashion from the luminosity , using the @xmath5@xmath204 relation of @xcite . as most of the maxbcg clusters", "do not have a significant flux ( or luminosity ) measurement , we can not calculate the individual cluster temperatures in this way . instead", ", each cluster temperature is approximated by the stacked spectral temperature from the appropriate richness bin , as shown in table  [ tab : lx750 ] .", "this ensures that our stacked cluster luminosities and individual luminosity estimates are on approximately the same footing .", "perhaps more importantly , changing the temperature does not change the 0.1 - 2.4 kev luminosities significantly  ( e.g. * ? ? ?", "for example , for the clusters with @xmath235 detections , if we follow the reflex iterative recipe to convert count rate to @xmath5 , rather than fixing the temperatures at the stacked values , the individual @xmath5 values change by @xmath236 .", "this is simply another example of how measuring temperatures with rosat is challenging .", "the 0.5 - 2.0 kev count rate is then converted to 0.1 - 2.4 kev luminosity using table 2 from @xcite with the equivalent hydrogen column density at the position of the cluster  @xcite . after converting to luminosity", ", a @xmath148-correction is applied using equation 4 from @xcite , which is a good approximation of table 3 in @xcite.z\\}^{1/2}$ ] . ]", "the @xmath148-corrections are not very large , at most 10% .", "we compared our method of calculating @xmath5 for individual clusters to the @xmath5 values obtained in the noras catalog ( see section  [ sec : xrayselect ] for details on the x - ray catalog matching ) .", "for the matched clusters the values are all consistent within errors with @xmath237 scatter , and a systematic offset of @xmath238 . the primary difference in our calculations", "is that the noras fluxes were calculated in an aperture obtained via growth curve analysis ( gca ) designed to obtain the best signal - to - noise for each cluster , while we use a fixed physical aperture .", "therefore , our fixed temperature and aperture provide an unbiased estimate of the cluster flux and luminosity even without _ a priori _ knowledge of extended cluster emission .", "individual cluster @xmath5 and @xmath6 values are shown in figure [ fig : linmix ] .", "we use linmix_err to estimate the power - law slope , normalization , and intrinsic scatter of the underlying distribution of the @xmath5@xmath6 relationship for the 955 richest clusters with @xmath239 .", "the best fit relation is : @xmath240 with an intrinsic scatter @xmath241 .", "the individual @xmath5 values are calculated within a fixed @xmath146 aperture , and thus this is to be compared to equation  [ eqn : lxn_750 ] .", "we test the robustness of this result by both splitting the input data into multiple independently fit subsamples and changing the richness threshold slightly .", "we find that the constraint on the scatter is robust , and the error bar is accurate .", "possible systematic biases in the constraint on @xmath221 are discussed in section  [ sec : coolcore ] .", "the scatter - corrected @xmath227@xmath1 relation is shown in figure  [ fig : linmix ] .", "the luminosities for the individual @xmath106 detections are plotted as solid circles , and the @xmath106 upper limits are plotted as empty circles . the typical error bar for detections", "is shown in the lower - right corner .", "contours showing the @xmath242 contours on the best - fit median relation are shown in dark gray , and the dashed ( dotted ) lines show the @xmath243 ( @xmath244 ) scatter constraints .", "the median relation as constrained by linmix_err has been converted to the equivalent mean relation by multiplying the normalization by @xmath229 , and is shown in light gray .", "the mean relation agrees well with the individual stacked bins ( squares ) , which are about @xmath231 brighter than the underlying median values .", "we must emphasize that we are measuring the _ observed _ scatter in the @xmath5@xmath6 relation as is appropriate to our catalog and stacking method .", "this scatter comes about not only due to intrinsic @xmath5 variations between the clusters ( including different morphologies and merging clusters ) , but is also due to point sources , cooling flows , and the projection of multiple clusters along the line of sight . for the @xmath245 richest clusters ( @xmath246 ) used in this exercise , the chance of projection is very low .", "we do not have any way to remove point sources or bright cool cores from individual clusters in an unbiased way .", "operationally , it is not relevant to our measurement whether the scatter is due to intrinsic @xmath5 variations or due to point source and cool - core contamination . to constrain the observed scatter ,", "we need only assume that the underlying distribution is approximately log - normal , which is consistent with our observations .", "@xcite measure the observed scatter in the @xmath5@xmath247 relation , where @xmath247 is a richness measure from the amplitude of galaxy - cluster correlation function  @xcite .", "they use 290 optically - selected clusters , of which 40 have significant detections in _ rosat_/pspc observations . using a bayesian maximum - likelihood fitting technique", ", they find a significant correlation between @xmath5 ( bolometric ) and @xmath247 , with a power - law slope of @xmath248 and an intrinsic scatter of @xmath249 .", "it must be noted that they introduce a prior weighting , @xmath250 , which gives more weight to models with lower scatter .", "again , it is difficult to directly compare our measurement of the slope and scatter to the @xcite measurements , due to different richness quantities and selection functions .", "a large intrinsic scatter in the @xmath5@xmath6 relation can account for optically - selected clusters that appear `` underluminous '' in x - rays as compared to their optical richness  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ) . in particular , @xcite present a set of `` abell x - ray underluminous '' ( axu ) clusters with masses determined from velocity dispersion measurements , that are significantly dimmer in x - rays than expected . in figure", "[ fig : linmix ] the seven axu clusters that match to maxbcg ( @xmath2 , @xmath233 ) are denoted with open squares .", "these are all significantly dimmer than the mean @xmath5@xmath6 relation , but are fully consistent with the observed scatter .", "velocity dispersion measurements of maxbcg clusters  ( b07 ) and comparisons to simulations  @xcite have shown that there is considerable mass - mixing in a given @xmath6 bin .", "the low - luminosity clusters could be from lower - mass halos that are picked up as moderate richness in the cluster finder .", "however , @xcite argue that the axu clusters have low @xmath5 relative to their mass .", "this could be due to biases in masses calculated from velocity dispersions with small samples  ( e.g. * ? ? ?", "* ) , or to large intrinsic scatter in the @xmath5@xmath13 relation  ( e.g. * ? ? ?", "each of these possibilities ( or a combination of the two ) are consistent with the observations .", "we further explore the @xmath5@xmath13 relation of maxbcg clusters in a companion letter  @xcite .", "we have reason to believe that the scatter is not fixed as a function of richness .", "for example , b07 find a significantly larger scatter in the velocity dispersion - richness ( @xmath251@xmath1 ) relation at low richness compared to high richness", ". however , the quality of our data for this exercise is not adequate to constrain the change in the scatter as a function of richness .", "therefore , for the remainder of the paper we adopt this nominal value of @xmath241 to correct the stacked ( mean ) @xmath0 to obtain the underlying median @xmath227 values .", "this essentially requires us to multiply all @xmath0 values by a factor of @xmath252 .", "we can compare the @xmath227@xmath1 relation for optically - selected maxbcg clusters to that from x - ray selected clusters from the literature .", "for this exercise we chose to compare to the noras  @xcite and 400 square degree  ( 400d : * ? ? ? * ) catalogs , both constructed from rosat data , and thus measured in the same energy range as used in this work .", "the noras catalog is an x - ray flux limited cluster catalog constructed from the rass photon maps in the northern sky .", "although it is known to be only @xmath55 complete , it provides a large sample of bright x - ray clusters that overlap the maxbcg survey region .", "the 400d catalog is a serendipitous x - ray flux limited cluster catalog constructed from pointed rosat pspc observations .", "the 400d catalog covers fields sampled from the whole sky , and thus only @xmath249 of the catalog overlaps the maxbcg survey region .", "our intention in this exercise is to obtain a baseline comparison of the x - ray and optical richness properties of x - ray selected clusters to optically - selected clusters .", "a somewhat different comparison of noras and maxbcg , focused on testing the completeness of maxbcg , was performed in @xcite .    to obtain a list of clean matches between noras or 400d clusters and maxbcg clusters , we require that the x - ray position and bcg position be matched within @xmath253 , and the redshift difference to be less than 0.05 .", "this ensures both that the probability of a false match is @xmath254 and that we are not selecting , e.g. , merging clusters where the x - ray and optical catalogs might utilize different deblending schemes .", "there are 89 noras clusters that match maxbcg , with a median redshift of @xmath255 , and 53 400d clusters that match maxbcg , with a median redshift of @xmath256 .", "we convert all x - ray luminosities to 0.1 - 2.4  kev in the rest frame with our adopted cosmology .", "note that the @xmath5 values obtained in noras and 400d have been corrected for aperture effects .", "however , with the adopted @xmath108 parameters ( @xmath257 for noras , and @xmath258 for 400d ) these corrections produce at most a @xmath237 offset in @xmath5 compared to the luminosities obtained in our scaled @xmath37 apertures .", "figure  [ fig : catmatches ] shows @xmath5 vs. @xmath6 for the x - ray selected clusters that meet our matching criteria .", "the solid circles represent noras clusters , and the empty squares represent 400d clusters .", "the luminosities of the 400d clusters are typically less than that for noras due to the deeper flux limit of the pointed rosat observations relative to the rass survey .", "the dashed line shows the median @xmath227@xmath1 relation , and the dotted lines show the @xmath243 scatter constraints .", "it is not surprising to note that x - ray selection picks out primarily the x - ray brightest clusters at a given richness ; nearly all the noras clusters lie above the median maxbcg relation .", "the deeper 400d survey selects a sample more representative of the optically - selected clusters , though it is still biased high .", "comparison to figure  [ fig : linmix ] is instructive . at every richness , the typical x - ray emission from clusters is significantly below what you might expect from x - ray selected catalogs .", "while this is true for both catalogs , it is especially apparent for noras clusters .", "although the 400d survey has the sensitivity to detect the richest x - ray dim clusters ( as in the lower - right corner of figure  [ fig : linmix ] ) , these clusters are quite rare .", "the limited overlap between the 400d survey region and the maxbcg survey region means that it is very unlikely that the 400d survey would contain one of these specific maxbcg clusters .", "there are a variety of systematic effects which may bias the @xmath227@xmath1 relation . in this section", "we address six : photometric redshift uncertainty , cluster centering errors , richness variation with redshift , bcg luminosity , point source contamination , and cool core clusters .", "the maxbcg cluster catalog is based on photometric data , and thus we only have photometric redshifts ( `` photo-@xmath31s '' ) for each of the clusters . while the photo-@xmath31s are relatively precise ( @xmath11 ) , we must investigate how using photometric redshift estimates might bias our stacking results .", "if a cluster is actually closer than its photometric redshift would suggest , it will be over - weighted in the stacking analysis . when it is more distant than it seems , it will be underweighted .", "even if the photo-@xmath31 errors are perfectly symmetric , the weighting factor , @xmath83 , is not . as a result ,", "uncertainty in photo-@xmath31 can introduce a bias in the luminosities .", "this bias can become significant if the photo-@xmath31 errors are large .", "the net impact of this effect is to make the clusters appear more luminous than they truly are .", "there is also a small volume effect caused by photo-@xmath31 errors . at a fixed redshift @xmath31", "there is a larger physical volume at a larger distance @xmath259 than at a smaller distance @xmath260 .", "because of this , more clusters are available at high redshift to scatter low than at low redshift to scatter high .", "this is similar to malmquist bias .", "as above , this effect will cause us to overestimate the cluster luminosities .", "in addition , the photometric redshift is used to calculate the cluster extraction radius .", "when a cluster is closer than its photo-@xmath31 implies , then the extraction radius used is too small , slightly reducing its implied luminosity .", "the opposite is true when the cluster is farther than its measured photo-@xmath31 .", "as there is very little cluster flux at large radii , and as this bias is roughly symmetric , we do not expect this effect to strongly bias the stacked @xmath5 .", "we run a monte carlo simulation to estimate the combined photo-@xmath31 bias as a function of richness . around @xmath261 of the bcgs in the maxbcg catalog used in this analysis", "have spectroscopic redshifts from the sdss dr5 spectroscopic catalog  @xcite . with the spectroscopic subsample", ", we measure the photo-@xmath31 offset ( @xmath262 ) distribution in each richness bin and three redshift bins .", "a monte carlo is performed to estimate the overall bias in the measured luminosity by comparing the `` true '' luminosities convolved with the @xmath263 distribution to the `` observed '' luminosities at the measured photometric redshifts .", "we find that the overall bias is small ( @xmath264 at 99% confidence limit ) , with the observed luminosities slightly overestimating the true luminosities .", "this bias factor does not vary with richness .", "we neglect this correction factor due to its small magnitude .      in our stacking exercise , we chose the bcg as the center of the cluster .", "for many clusters with individual x - ray observations and a bright bcg , this is coincident with the x - ray center .", "however , there are several types of clusters in which the bcg is not consistent with the center of the x - ray emission .", "these include merging clusters with disturbed morphologies ; clusters without a dominant central galaxy for unambiguous detection by the maxbcg algorithm ; and a few ( @xmath265 ) previously identified strong cooling flow clusters .", "the central galaxies in these clusters host bright agn and exhibit strong star formation , moving their colors off the red sequence and excluding them from the maxbcg cluster finding algorithm .", "we do not expect centering biases to be a significant problem for our calculation of the mean x - ray luminosity for the following reasons .", "first , the x - ray luminosity is proportional to the electron density @xmath266 , and thus most of the luminosity comes from the central region . with our large apertures ( @xmath267 )", "as long as the core of a cluster is within the aperture , the photon counting exercise will include most of the cluster photons .", "second , we have reason to believe from simulations that most of the richer clusters are well centered  ( j07 ) .", "this is not to say that the cluster centering will not have any effect : as shown in section  [ sec : profiles ] , the @xmath108-model parameter is highly dependent on the centering distribution .    to confirm this , we perform some simple tests of de - centering the clusters . if we randomly offset all the cluster positions with a 2d gaussian with an rms of @xmath268  ( comparable to the centering distribution modeled in j07 ) , the mean @xmath0 values decrease by @xmath269 , and less than @xmath106", "this rms value was chosen because , as is seen in figure  [ fig : ptmatchhist ] , the vast majority of the x - ray cluster matches are within @xmath253 . therefore , we do not consider cluster miscentering to contribute a significant bias to our x - ray luminosity calculation .", "we investigate whether the observed @xmath0 at fixed richness changes with redshift .", "modest evolution in @xmath0 at fixed _ mass _ is expected .", "if the clusters are evolving in a self - similar manner , then we expect the higher redshift clusters to be more luminous than the lower redshift clusters , due to the fact that the universe was more dense at higher redshift .", "the expectation is that @xmath270 , where @xmath271 is the critical density of the universe at redshift @xmath31  @xcite .", "therefore , clusters at a given _ mass _ at a redshift @xmath272 should be @xmath273 brighter than similar clusters at a redshift of @xmath274 .", "to study possible variation of @xmath0 at fixed richness , we split the cluster sample into three redshift bins , containing the bottom 25% , middle 50% , and top 25% of clusters in our sample .", "the low redshift bin ranges from @xmath275 with a median redshift of @xmath276 ; the middle redshift bin ranges from @xmath277 with a median redshift of @xmath41 ; and the high redshift bin ranges from @xmath278 with a median of @xmath279 . to obtain increased signal - to - noise in these bins , we combine richness bins @xmath63 and @xmath62 ; @xmath65 and @xmath64 ; and @xmath280 and @xmath66 .", "these wider bins are the same that were used in the lensing analysis of s07 .", "figure  [ fig : splitzplot ] shows the stacked @xmath0 as a function of @xmath1 for the three different richness bins .", "the dashed line shows the mean relation ( equation  [ eqn : lxn_r200 ] ) for all clusters from section  [ sec : meanrelation ] .", "the high redshift clusters ( diamonds ) are significantly more luminous than the low redshift clusters ( circles ) .", "we parameterize the variation with a factor of @xmath281 , and fit all the redshift and richness bins simultaneously with a model of the form : @xmath282 where @xmath41 , the median redshift of the cluster catalog .", "this results in a good fit ( @xmath283 ) with the best - fit parameters : @xmath284 ; @xmath285 ; @xmath286 . note that @xmath287 and @xmath108 are consistent with the mean relation in equation  [ eqn : lxn_r200 ] .", "the redshift variation parameter , @xmath288 , is quite large , and shows that the high redshift bin ( @xmath279 ) is almost twice as bright as the low redshift bin ( @xmath276 ) .", "this is significantly in excess of the self - similar prediction .", "the challenge is to determine the origin of this redshift dependent shift in @xmath0 at fixed richness .", "the photo-@xmath31 errors described in section  [ sec : photoz ] may account for some of this effect .", "although the absolute photo-@xmath31 errors are better than @xmath289 at all redshifts , the _ relative _ photo-@xmath31 errors are larger at low redshift than at high .", "therefore , the impact of photo-@xmath31 bias is larger at low redshift than at high redshift .", "we repeated the monte carlo simulation described in section  [ sec : photoz ] to measure the photo-@xmath31 bias from each redshift bin .", "the additional redshift bias is @xmath290 ( 99% confidence limit ) in the low redshift bin , and @xmath291 ( 99% confidence limit ) in the high redshift bin .", "the photo-@xmath31 bias therefore remains smaller than the observed variation in @xmath5 , although the effect is in the same sense as the variation we observe , and is a contributing factor .", "a more likely possibility is redshift dependent variation of @xmath6 at fixed mass .", "if our high redshift clusters have systematically smaller @xmath6 at fixed mass , this would have the effect of shifting the high redshift points in figure  [ fig : splitzplot ] to the left . in order to constrain this bias ,", "we must first factor out the effect of the self - similar evolution of @xmath0 at fixed mass , using the expected redshift dependence @xmath292 as described above .", "we then re - fit the data to equation  [ eqn : zgamma ] to calculate the excess variation in the @xmath0@xmath1 relation that may be attributed to variation of @xmath6 at fixed mass .", "this results in a best - fit @xmath293 , implying a fractional decrease in @xmath6 of @xmath294@xmath295 from our lowest redshift bin ( @xmath296 ) to our highest redshift bin ( @xmath297 ) .", "this is consistent with the redshift dependent variation in the velocity dispersion ", "optical richness relation measured by b07 .", "however , we must note that redshift dependence in the observed lensing shear , @xmath298 , is significantly smaller  ( s07 ) .", "unfortunately , none of these approaches is yet able to confidently determine the nature of the observed variation .", "for example , this effect could be caused by incorrect handling of the 0.4  @xmath18 and color cuts that determine which galaxies are included in the richness estimate .", "it might also be `` true '' evolution , such that clusters have fewer red - sequence galaxies brighter than 0.4 @xmath18 at fixed mass at higher redshifts .", "further work , including improving the richness estimates , is in progress to constrain the nature of this evolution .", "we now investigate the effect bcg @xmath30-band luminosity ( @xmath299 , see @xcite for details ) has on the mean x - ray luminosity @xmath0 .", "simulations and semi - analytical modeling predict that dark matter halos formed at early times have brighter bcgs and lower richness than those which form late  @xcite .", "thus , we expect that the x - ray luminosity should be correlated with @xmath299 at fixed richness .", "furthermore , optical and x - ray observations of individual nearby clusters show that more luminous bcgs are correlated with more massive halos with higher x - ray temperature and luminosity  ( e.g. * ? ? ?", "this trend has already been seen in the mean velocity dispersion analysis of maxbcg clusters , as the clusters with more luminous bcgs had significantly larger velocity dispersions than clusters with similar richness  ( b07 ) .    in an exercise similar to that performed in the previous section , we split each wide richness bin into three bins of @xmath299 .", "as @xmath299 , unlike redshift , is correlated with @xmath6 , we are unable to use the same @xmath299 split for each @xmath6 bin .", "therefore , we sort the clusters in each richness bin by @xmath299 , and split the sample into the top 25% , middle 50% , and bottom 25% , and then restack .", "the results are shown in figure  [ fig : splitbcgplot ] . it is readily apparent that the clusters with the most luminous bcg in a given richness bin ( diamonds ) are significantly more luminous in x - rays than other clusters in the bin .", "this effect is most dramatic at lower richnesses .", "this is expected , as the bcg is more dominant in low richness clusters than high richness clusters  ( e.g. * ? ? ?", "it is also notable that there does not appear to be a significant difference between the @xmath0 for the low @xmath299 and middle @xmath299 bins .", "we parameterize the effect of @xmath299 on @xmath0 using a similar procedure as performed for the redshift variation .", "the richness and @xmath299 bins are fit simultaneously with a model of the form : @xmath300 this results in an adequate fit ( @xmath301 ) with the best - fit parameters : @xmath302 ; @xmath303 ; @xmath304 .", "although the normalization is the same as in equation  [ eqn : lxn_r200 ] , the slope is significantly more shallow .", "this is caused by the correlation between @xmath299 and @xmath6 .", "after taking this into account , the mean relation shown here is fully consistent with that calculated without splitting the bins by @xmath299 .", "the reason the @xmath145 of the fit is relatively high is because it does not appear that @xmath0 is smooth function of @xmath305 : the effect of @xmath299 is not symmetric , and only tends to boost the x - ray luminosity of clusters with relatively bright bcgs .", "it is not entirely clear whether this effect is due to an imperfect cluster finder , our crude richness definition , or due to real cluster physics .", "clusters with brighter , more dominant bcgs are more likely to be correctly centered and thus might be more luminous in x - rays .", "however , as we showed in section  [ sec : centering ] , the effect of decentered clusters on @xmath0 is much smaller than the effect we see due to @xmath299 .", "meanwhile , our richness definition , @xmath6 , is a count of red - sequence galaxies brighter than @xmath306 .", "it is thus not surprising that a poor cluster with an exceptionally bright bcg might be more massive and more x - ray luminous than a cluster at a similar richness with a more typical bcg .", "finally , it is possible that these clusters are associated with halos that formed early , allowing many of their member galaxies to merge into a very large bcg  @xcite . deeper targeted x - ray observations of specific clusters will be required to determine if we can use @xmath299 and other optical properties to constrain the age of its dark matter halo .      when calculating the projected x - ray luminosity from the hot intracluster medium , there is the possibility of contamination due to x - ray point sources .", "some point sources , such as cluster agn , are associated with the cluster , but in general their luminosities are not tightly correlated with that of the icm .", "other point sources that are chance coincidences , such as foreground stars and background quasars , might also boost the apparent luminosity .", "the problem of point source contamination is exacerbated by the broad psf of the rass survey .", "due to the large psf , it is difficult to accurately excise point sources from the higher redshift clusters , even when their positions are known . for the stacking exercise ,", "foreground and background sources are not a significant problem . assuming the positions of these point sources are uncorrelated with the cluster positions , the contribution from these sources is subtracted with the background estimation , as is demonstrated by our stack of random points in figure  [ fig : imageplots ] . at the same time , these point sources , both those associated with the clusters and those that are chance projections , may increase the observed scatter . in general", ", we note that @xmath5 values calculated from rass are not corrected for point source contamination  ( e.g. * ? ? ?", "* ; * ? ? ?", "we wish to perform a first - order check on possible contamination from cluster agn , the dominating point sources that might bias the observed @xmath0 .", "@xcite have recently performed a detailed survey of the distribution of agn in galaxy clusters for eight moderate redshift clusters , comparable to the redshift range of the maxbcg catalog .", "they find that moderately bright agn ( @xmath307 ) associated with bright cluster members ( @xmath308 ) make up around @xmath309 of the total cluster member population .", "this implies that many of the richest maxbcg clusters contain at least one moderately bright agn .", "these clusters have an extremely bright icm ( @xmath310 ) , and thus the fraction of the luminosity from cluster agn will , on average , be very small .", "if we extrapolate the same agn fraction to our poorest clusters ( @xmath311 , perhaps one in ten will host a moderately bright agn . although these agn might have a luminosity comparable to the icm for these @xmath312 of poor clusters , this small fraction should not strongly bias the stacked average .    as a simple check for contamination in the x - ray signal from cluster agn", ", we match maxbcg cluster member galaxies to radio sources from the first survey  @xcite with a match radius of @xmath313 . by matching random galaxies from sdss dr5 to the first sources", ", we estimate the purity of the matches to be @xmath314 .", "the first survey covers approximately the same footprint as the sdss , down to a typical flux limit of @xmath315 at 1.4 ghz .", "using 1.4 ghz radio detections as a proxy for agn activity has the advantage that the radio spectrum is often flat , and thus the 1.4 ghz flux is not a strong function of redshift .", "the disadvantages are that only 10% of x - ray bright quasars and agn are radio loud , and that there is a large scatter between x - ray and radio luminosities of these objects . for example", ", @xcite noted that for the same small set of clusters none of the x - ray selected agn are radio loud  @xcite .", "although we have an incomplete selection function , we can still constrain the contamination from cluster agn . if the mean @xmath0 is significantly boosted by cluster agn", ", then we would expect the clusters which match radio sources to be relatively bright .", "this effect would be greatest at low richness and low @xmath0 when the icm is not as hot or bright .", "for each richness bin , we matched the maxbcg cluster members within @xmath146 of the bcg to sources in the first survey .", "table  [ tab : radiomatch ] shows the number of clusters in each bin that have members matched with the first catalog .", "more detailed work is in progress in cross - correlating the maxbcg catalog with radio sources to explore cluster agn feedback and related issues ( see also @xcite ) .", "ccc @xmath61 & 45 & 82% + @xmath62 & 91 & 65% + @xmath63 & 125 & 62% + @xmath64 & 183 & 54% + @xmath65 & 284 & 45% + @xmath66 & 420 & 37% + @xmath67 & 332 & 30% + @xmath68 & 1330 & 25% + @xmath69 & 1341 & 18% +    we wish to compare how the radio selected subset of clusters compares to a randomly selected subset of clusters of similar richness . in each richness bin , we ran 10000 trials by sampling ( with replacement ) from the entire set of clusters in the bin , sampling the same number of clusters that match radio sources . as a quick estimate of the mean luminosity , rather than performing the full stacking analysis and spectral fits , we calculated the stacked and weighted , background - subtracted hard channel ( 0.5 - 2.0 kev ) counts in a fixed @xmath146 aperture .", "the resulting count rate was converted to @xmath0 using the method described in section  [ sec : scatter ] .", "we can then compare the mean @xmath0 of the clusters which match radio sources to the typical values of @xmath0 that we would expect when drawing that same number of clusters from the entire sample .", "the results are shown in figure  [ fig : radiomatch ] . in each panel", "the histogram shows the distribution of @xmath0 values obtained from the resampling , and the vertical dashed line marks the mean @xmath0 of the radio - matched clusters .", "we do not see any significant bias as a function of richness .", "the only bin in which the radio matches appear to be outliers is the @xmath316 richness bin .", "this is due to random chance , in that the clusters in this richness bin with the highest @xmath5 also happen to match radio sources ; although there is no obvious indication in the rass data that there is point source contamination on these clusters , further follow - up with _ chandra _ or _ xmm / newton _ would be required to clarify this .", "there is much evidence that cool core clusters increase the scatter in x - ray cluster properties . by excluding cluster cores in high resolution imaging ,", "many x - ray parameters become more tightly correlated  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "the broad rosat psf , combined with the fact that most of our clusters do not have individual x - ray detections , makes it impossible to exclude cluster cores in our stacking analysis . however , a few of the brightest clusters ( @xmath317 ) at moderate richness ( @xmath318 ) are well known cool - core clusters .", "we thus investigate what effect cool core clusters have on our estimate of the median @xmath227 and scatter @xmath221 in the @xmath227@xmath1 relation .", "there does not exist an unbiased volume - limited catalog of cool core x - ray clusters which can be fairly compared to the maxbcg catalog .", "@xcite used rosat pointed observations of an x - ray flux limited catalog  @xcite to estimate the central cooling time of 55 nearby clusters .", "two of these clusters are in the maxbcg catalog ( a1689 , a2244 ) and exhibit characteristics of a cool core ( @xmath319 ) .", "@xcite used _ chandra _ observations of the higher redshift ( @xmath320 ) bcs clusters to systematically estimate the central cooling time of bright x - ray clusters .", "we identify 7 maxbcg clusters ( a750 , a963 , a1835 , rxj 2129.6 + 0005 , z2701 , z3146 , z7160 ) from this sample with cooling times @xmath319 which we mark as cool core clusters .", "finally , we identify clg  j1504 - 0248  @xcite as maxbcg cluster with a cool core .", "it should be noted that most of these clusters have a moderate richness ( @xmath321 ) and are in the brightest tail of the @xmath5 distribution for their richness .", "furthermore , the central galaxies of these clusters tend to show strong @xmath322 emission in sdss spectroscopy .", "in fact , for a few of these clusters the true bcg is not in the list of cluster members because the strong @xmath322 emission changes the apparent galaxy color , making it inconsistent with the red - sequence used in the cluster finding algorithm .    after excluding these known cool core clusters ( which might bias the @xmath227@xmath6 relation ) we recalculate the stacked mean relationship as well as the median relationship with scatter . when calculating the median relationship with scatter , as in section  [ sec : scatter ] , the normalization and slope shift by @xmath323 , as these values are not strongly affected by outliers .", "however , the intrinsic observed scatter decreases to @xmath324 , which is a @xmath325 shift .", "it is not surprising that @xmath221 decreases : we are deliberately removing the brightest clusters .", "meanwhile , the stacked relation is also slightly affected by these bright outliers , and after cutting the known cool core clusters the normalization of the stacked relation decreases by @xmath326 . after following the previously described prescription for converting the stacked mean relation to a scatter - corrected median relation", ", we find that removing the known cool core clusters decreases the normalization by @xmath209 .", "therefore , it does not appear that the known , very x - ray luminous , cool core clusters , are significantly biasing our relation .", "we reiterate that the intrinsic scatter we measure is the _ observed _ scatter that takes into account _ all _ x - ray observations of the maxbcg clusters , regardless of x - ray morphology of the clusters .", "thus , it is likely that some portion of this observed scatter is caused by unidentified cool core clusters in the sample .", "b07 have measured the median velocity dispersion ( @xmath328 ) as a function of @xmath1 for the maxbcg catalog .", "the richness bins used to measure the stacked velocity dispersions , as well as the scatter in the @xmath12@xmath6 relation , are the same as those used in this work . however , in order to measure velocity dispersions , only clusters that have the bcg and at least one additional member galaxy in the sdss dr5 spectroscopic subsample  @xcite were used .", "the spectroscopic subsample preferentially selects bright / nearby galaxies , and thus the selection of clusters used in b07 ( @xmath329 , with @xmath330 ) is slightly different than that for the full maxbcg catalog ( @xmath331 , with @xmath332 ) . as discussed in section  [ sec : redshift ] , the lower redshift clusters have a smaller @xmath0 than the higher redshift clusters .", "we therefore re - run our stacking procedure on exactly those clusters used in b07 , and use the known spectroscopic redshifts of the bcgs rather than the photometric redshifts .", "this ensures that , for this exercise at least , we are not affected by any possible photometric redshift biases , which otherwise could be significant for the nearest clusters ( @xmath333 ) where the relative photo-@xmath31 error ( @xmath334 ) is quite large . table  [ tab : lxsig ] shows the richness bins used , as well as the mean @xmath0 ( this work ) and median @xmath327 ( from b07 ) for the nine richness bins .", "the slope of the @xmath0@xmath1 relation is nearly identical for the spectroscopic subsample as reported in section  [ sec : meanrelation ] , but the normalization is lower by @xmath312 , due to the different redshift selection .", "ccccc @xmath61 & 83.9 & 1.69 & @xmath335 & @xmath336 + @xmath62 & 58.4 & 1.45 & @xmath337 & @xmath338 + @xmath63 & 44.7 & 1.32 & @xmath339 & @xmath338 + @xmath64 & 35.9 & 1.22 & @xmath340 & @xmath341 + @xmath65 & 28.6 & 1.10 & @xmath342 & @xmath343 + @xmath66 & 22.7 & 1.02 & @xmath344 & @xmath345 + @xmath67 & 19.0 & 0.94 & @xmath346 & @xmath347 + @xmath68 & 14.0 & 0.82 & @xmath348 & @xmath349 + @xmath69 & 9.9 & 0.73 & @xmath350 & @xmath351 +    we take the mean @xmath0 values and corrected them for scatter as described in section  [ sec : scatter ] to obtain @xmath227 .", "figure  [ fig : lxsig ] shows the median @xmath227@xmath327 relation for maxbcg clusters with richness @xmath352 . each data point ( solid circles ) is obtained from the same richness bin and same cluster selection .", "the dark gray band shows the best - fit relation ( @xmath353 ) from clusters in the spectroscopic subsample of the maxbcg catalog , with the following functional form : @xmath354 the scaling relation predicted by self - similar cluster evolution is that @xmath355  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ) , which is consistent with our observations .", "we note that if we could convert our 0.1 - 2.4  kev luminosities to @xmath202 , this would steepen our relation , as the clusters with higher velocity dispersions also have a higher temperature and larger @xmath356 ratio .", "as mentioned previously , such a conversion is not practical for the stacking exercise using rosat data .", "the slope of our 0.1 - 2.4 kev @xmath227@xmath327 relation is consistent with that measured for reflex clusters  @xcite .", "we compare this median relation to that obtained previously for individual x - ray selected clusters .", "@xcite calculated the velocity dispersions of reflex clusters by cross - correlating the reflex catalog with the two - degree field galaxy redshift survey  ( 2df ; * ? ? ?", "figure  [ fig : lxsig ] shows 39 individual reflex clusters ( empty circles ) from hilton , et al .", "( 2007 , private communication ) after converting the velocity dispersions to the cluster rest frame . in the upper - left corner", "the empty circle shows the typical errors in 0.1 - 2.4 kev @xmath5 and @xmath12 for the reflex/2df clusters .", "@xcite found the slope of the x - ray selected @xmath5@xmath12 relation to be @xmath357 , consistent with our observations .", "however , the individual x - ray selected clusters tend to have a slightly higher ( @xmath326 ) velocity dispersion at a given x - ray luminosity compared to the stacked maxbcg clusters .", "this is possibly caused by the difficulty in measuring velocity dispersions for individual clusters .", "contamination from non - virialized galaxies and interlopers will more likely result in an overestimate of the velocity dispersion rather than an underestimate .", "by stacking observations from the _ rosat _ all - sky survey , we measure the mean x - ray luminosity as a function of richness for optically - selected clusters from the maxbcg catalog . with a large number of clusters in each bin , the stacking exercise has the power to probe to much lower flux limits than are possible for x - ray selected surveys .", "thus , a highly pure and complete volume - limited catalog with clusters selected by their optical properties can be used to measure the mean @xmath0 .", "we find that @xmath0 scales with optical richness , @xmath6 , with a simple power - law form over two orders of magnitude in @xmath0 .", "our results are similar to those obtained in dkm07 by stacking rass observations of clusters identified in 2mass .", "however , we find that stacked temperature measurements from _ rosat _ are significantly biased for higher temperature clusters .", "therefore , we restrict our analysis to luminosities calculated in the 0.1 - 2.4  kev band rather than extrapolating to calculate bolometric luminosities . furthermore , we have shown that the @xmath108-model fits to the stacked radial profiles of moderate redshift clusters are dominated by a combination of the broad _ rosat _ psf and the offset distribution between maxbcg clusters and correlated x - ray sources . on the other hand , calculating x - ray luminosity in the 0.1 - 2.4 kev band , which relies primarily on counting photons , is not strongly biased by spectral fits or profile fits . as @xmath5 scales with the square of the density , as long as the core of a cluster is within the stacking aperture , we can obtain a high reliability estimate of the mean cluster luminosity .    by measuring the x - ray flux at the positions of individual maxbcg clusters , we are able to constrain the observed scatter in the @xmath5@xmath6 relation . assuming a log - normal distribution of @xmath5 as a function of @xmath6 , we find @xmath241 .", "this scatter is quite large , and thus it is necessary to correct for the scatter when calculating the median of the underlying distribution @xmath227 .", "the richness measure we have used in this work , @xmath6 , is simply the count of red - sequence galaxies within a scaled @xmath37 aperture , but nevertheless is a good proxy for @xmath5 on average .", "it is likely that improved richness measurements can be made . from the evidence presented here , it seems likely that these new richness estimators will include information about bcg luminosity , and will more carefully control variations in cluster membership with redshift .", "the large scatter in the @xmath5richness relation does have some significant effects on cluster selection .", "first , x - ray selection with a high flux limit , as with the noras catalog , tends to pick out the clusters with the highest @xmath5 at a given richness .", "second , the large scatter in the @xmath5@xmath6 relation means that there are a significant number of optically selected clusters that are seemingly `` underluminous '' .", "optically selected clusters will often appear relatively dim compared to x - ray selected clusters at a similar richness .", "the goal of cluster selection and identification is to find a low scatter proxy for halo mass .", "both x - ray and optical selection techniques suffer from different limitations .", "x - ray flux limited surveys can only find the brightest x - ray clusters , and are thus limited to a combination of the brightest and nearest clusters . furthermore , there may be significant scatter between x - ray luminosity and halo mass  @xcite .", "optical surveys such as maxbcg may be volume limited out to moderate redshift , but optical richness also has a large scatter with respect to x - ray luminosity and mass  ( b07 ) .", "the development of more precise mass proxies , including @xmath204 , the compton y parameter @xcite , @xmath358  @xcite , and potentially more sophisticated optical richness estimates may allow selection of cluster catalogs more closely approximating the mass - limited catalog we aspire to .    by combining our present analysis with the velocity dispersion measurements of b07", ", we have measured the median @xmath227@xmath327 relation for the maxbcg clusters .", "the relation has a slope of @xmath359 , consistent with the prediction of self - similar cluster evolution .", "previous measurements of the @xmath5@xmath12 relation have been complicated by both selection effects and a large scatter , as the derived slope depends strongly on the fitting technique used . by stacking clusters from a volume - limited catalog and correcting for scatter , both @xmath227 and @xmath327", "are much better constrained , allowing a robust calculation of the relation . in our analysis", ", we do not see any evidence of a break in the @xmath227@xmath327 relation at the poor cluster ( group ) scale , as has been hinted at previously  ( e.g. * ? ? ?", ". this might be due to the fact that the maxbcg cluster finder works with red - sequence clusters , and does not find groups with a large blue galaxy fraction .", "we can also compare @xmath0 as determined from this method to weak lensing masses determined from the same maxbcg cluster catalog  ( s07 , j07 ) .", "determining the @xmath5@xmath13 relation is not only important for understanding cluster physics , but for calibrating the selection function of x - ray clusters used as cosmological probes  ( e.g. * ? ? ?", "we have made this measurement in a companion letter  @xcite .", "in addition , the comparison of the x - ray luminosities , weak lensing masses , and velocity dispersions can be used in conjunction with the number function of maxbcg clusters to constrain cosmological parameters  ( e.g. * ? ? ? * ) .", "this work is ongoing .", "refined measurements of the scatter , point source contamination , and confirmation of catalog purity will require deeper pointed x - ray observations .", "we are currently investigating serendipitous maxbcg cluster observations by deeper pointed _ rosat_/pspc observations , as well as _", "xmm / newton _ and _ chandra _  ( e.g. * ? ? ?", "unfortunately , most of the serendipitous cluster observations are of relatively poor clusters , which are much more numerous .", "a targeted campaign of a representative sample of maxbcg clusters that have not previously been known to be x - ray bright will be essential to both test our measurement of the scatter in @xmath5@xmath6 and to estimate the fraction of cluster flux that is contaminated by point sources .", "this will be essential to enable future inexpensive deep optical cluster surveys such as des  @xcite to use optical properties of clusters to estimate which clusters would be most useful to follow - up with targeted x - ray observations .", "e. rykoff and t. mckay are pleased to acknowledge financial support from nsf ast-0206277 and ast-0407061 , and the hospitality of the michigan center for theoretical physics .", "mrb acknowledges the support of the michigan space grant consortium .", "aee acknowledges support from nsf ast-0708150 .", "we also wish to thank the anonymous referee for many helpful comments .", "funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .", "department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .", "the sdss web site is http://www.sdss.org/.    the sdss is managed by the astrophysical research consortium for the participating institutions .", "the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .                                                                                                                                                                , e. et  al .", "1987 , in presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol .", "733 , soft x - ray optics and technology ; proceedings of the meeting , berlin , federal republic of germany , dec .", "8 - 11 , 1986 , bellingham , wa , society of photo - optical instrumentation engineers , volume 733 , 1987 , p. 519 . ,", "koch & g.  schmahl , 519+                                          , w. , boller , t. , englhauser , j. , freyberg , m. , & supper , r. 2001 , in astronomical society of the pacific conference series , vol .", "225 , virtual observatories of the future , ed .", "r.  j. brunner , s.  g. djorgovski , & a.  s. szalay , 234+"], "abstract_text": ["<S> determining the scaling relations between galaxy cluster observables requires large samples of uniformly observed clusters . </S>", "<S> we measure the mean x - ray luminosity  optical richness ( @xmath0@xmath1 ) relation for an approximately volume - limited sample of more than 17,000 optically - selected clusters from the maxbcg catalog spanning the redshift range @xmath2 . by stacking the x - ray emission from many clusters using _ rosat _ all - sky survey data , </S>", "<S> we are able to measure mean x - ray luminosities to @xmath310% ( including systematic errors ) for clusters in nine independent optical richness bins . in addition </S>", "<S> , we are able to crudely measure individual x - ray emission from @xmath4 of the richest clusters . assuming a log - normal form for the scatter in the @xmath5@xmath6 relation , we measure @xmath7 at fixed @xmath6 . </S>", "<S> this scatter is large enough to significantly bias the mean stacked relation . </S>", "<S> the corrected median relation can be parameterized by @xmath8 , where @xmath9 and @xmath10 . </S>", "<S> we find that x - ray selected clusters are significantly brighter than optically - selected clusters at a given optical richness . </S>", "<S> this selection bias explains the apparently x - ray underluminous nature of optically - selected cluster catalogs . </S>"], "labels": null, "section_names": ["introduction", "input data", "x-ray analysis", "mean @xmath0@xmath1 relation", "biases in the @xmath227@xmath1 relation", "the luminosityvelocity dispersion relation (@xmath227@xmath327)", "summary"], "sections": [["galaxy clusters , as the largest peaks in the cosmic density field , play an important role in astrophysics and cosmology ( e.g. * ? ? ?", "* ; * ? ? ?", "structure formation theory , realized in large scale n - body simulations , makes robust predictions for cluster space density and clustering within various cosmological models . since clusters are also the most observationally accessible features of large scale structure , they provide an opportunity to place strong constraints on both cosmological parameters and the growth of structure .", "the great challenge of cluster cosmology lies in confidently relating the dark matter halos we can robustly predict to the baryonic structures we observe .", "substantial work is being done to close the gap between theory and observations from both sides .", "on the theory side , numerical simulations of ever increasing complexity and resolution provide new insights into the evolution of baryons within clusters , and to elucidate the connection between cluster galaxies and dark matter substructure @xcite .", "observationally , much work is being done to assemble large samples of clusters detected and observed in a wide variety of ways .", "this is possible because galaxy clusters provide a rich suite of observables .", "optical light is emitted by individual cluster galaxies as well as intracluster stars .", "x - rays are emitted by both the hot intracluster medium ( icm ) and agn within cluster galaxies .", "this same hot icm scatters microwave background photons passing through the cluster , distorting their spectrum .", "finally , the total projected mass distribution of the cluster produces weak and sometimes strong lensing distortions in the images of background galaxies .", "each of these observables presents an opportunity to detect clusters and measure their properties . when combined , they allow us to cross - check our understanding of cluster physics in a variety of ways .", "clusters were first detected as early as the 18th century as anomalous groupings of similarly bright galaxies @xcite .", "optical surveys , which are relatively inexpensive , have long provided the largest cluster catalogs , primarily because of their ability to detect objects with relatively low mass thresholds . while successful in identifying thousands of clusters @xcite and providing the first evidence for dark matter @xcite , early optical detection was plagued by projection of galaxies along the line of sight @xcite .", "precise ccd photometry has enabled searches for galaxies clustered in space , brightness , and color  ( for a review of optical selection , see * ? ? ? * ) , substantially reducing the problems of projection .", "these more recent optical surveys also naturally provide accurate photometric redshifts .    with", "the advent of x - ray satellites , detection of thermal emission from the hot icm became possible @xcite .", "because x - ray emission depends on the square of the density , this provides a higher contrast that is largely immune to projection effects .", "but when spatial resolution is low , contamination from non - thermal x - ray sources can be a difficult problem", ". these can be point sources like agn , the non - thermal emission from a cooling core , or ongoing merger activity that has thrown a cluster out of thermal equilibrium . over the last several decades", ", x - ray surveys have been used to assemble a large number of cluster catalogs @xcite .", "much of our understanding of cluster physics today derives from these x - ray selected catalogs .", "substantial effort has gone into comparing optically and x - ray selected catalogs @xcite , and to comparing optical and x - ray properties of clusters to their weak lensing and sz signals @xcite .", "comprehensive comparisons of optical and x - ray properties of clusters have been hampered by the lack of large samples that are uniformly observed in both passbands .    in this paper", "we describe measurements of the x - ray properties of the largest publicly available optically - selected cluster sample : the maxbcg catalog @xcite .", "this approximately volume - limited cluster catalog spans the redshift range from @xmath2 .", "uniform optical photometry and relatively precise photometric redshifts ( @xmath11 ) for all these clusters are available from the same sloan digital sky survey  ( sdss : * ? ? ?", "* ) data from which the clusters were selected .", "x - ray observations of all these maxbcg clusters are available from the _ rosat _ all - sky survey  ( rass : * ? ? ?", "while rass exposures are too shallow to allow significant individual detections of every maxbcg cluster , they provide precise measurements of the mean x - ray luminosity ( @xmath0 ) as a function of optical richness and redshift .", "in addition , the low signal - to - noise measurements of x - ray emission from individual clusters can be used both to confirm the measurement of mean x - ray emission obtained by stacking and to provide estimates of the scatter in the optical richness  x - ray luminosity relation .", "the maxbcg catalog has been studied in a variety of complementary ways .", "for example , both dynamical  ( * ? ? ? * henceforth b07 ) and weak lensing  ( * ? ? ?", "* ; * ? ? ?", "* s07 , j07 ) observations of these clusters have been extracted from sdss data .", "these earlier observations can be combined with the mean @xmath0 measurements presented here to provide further insight into cluster physics .", "the @xmath0@xmath12 relation inferred for maxbcg clusters is described here , while the corresponding @xmath0@xmath13 relation obtained from weak lensing is discussed in a companion letter  @xcite .", "an analysis similar to that reported in this paper was performed for the nir selected 2mass cluster catalog by ( * ? ? ?", "* henceforth dkm07 ) .", "the 2mass nir flux - limited catalog has @xmath14 nearby ( @xmath15 ) groups and clusters ranging in mass from @xmath16 , selected with a matched - filter algorithm  @xcite .", "dkm07 are the first to take a large optically - selected cluster catalog and measure the mean ( stacked ) x - ray properties using rass , rather than simply cross - correlating optically - selected and x - ray - selected clusters .", "they find that the x - ray luminosity of 2mass clusters scales with optical richness ( their @xmath17 , the number of galaxies brighter than @xmath18 within @xmath19 ) .", "in addition , they derive x - ray temperatures and hydrostatic masses for the stacked 2mass clusters .    in section [ sec : input data ]", ", we briefly review the sdss data , maxbcg catalog , and rass data on which this study is based .", "section [ sec : x - ray analysis ] describes our measurement of x - ray luminosities for individual clusters , as well as our methods for determining the mean x - ray luminosity of a set of clusters with similar richness .", "we describe the mean relation of x - ray luminosity as a function of richness , scatter in this relation , and the underlying median @xmath20 relation in section [ sec : meanrelation ] .", "section [ sec : biases ] discusses several possible sources of systematic bias in these results . in section [ sec : lsig ] we combine these results with dynamical measurements of maxbcg clusters to produce a measurement of the @xmath21 relation .", "conclusions and some discussion of future work are presented in section [ sec : summary ] . throughout this work we use a @xmath22cdm cosmology with @xmath23 and @xmath24 ."], ["the measurements described here are based on two wide - area imaging surveys ; sdss and rass .", "galaxy clusters are selected from the sdss five - band imaging data using a red sequence selection method .", "sdss data also allow measurement of cluster redshifts and richnesses .", "x - ray emission from these clusters is then measured by from the rass photon maps . in this section", "we describe briefly the sdss imaging data , galaxy cluster selection and calibration , and rass input data .", "optical data for this study are drawn from sloan digital sky survey : a combined imaging and spectroscopic survey of 10@xmath25 deg@xmath26 in the north galactic cap and a smaller region in the south .", "the imaging survey was carried out in drift - scan mode in five sdss filters ( @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 ) to a limiting magnitude of @xmath32 @xcite .", "photometric errors are typically limited at bright magnitudes by systematic uncertainties at the 3% level .", "the spectroscopic survey targets both a `` main '' sample of galaxies with @xmath33 and a median redshift of @xmath34 @xcite and a `` luminous red galaxy '' sample @xcite which is approximately volume limited out to z=0.38 . for more details of the sdss", "see @xcite and @xcite .", "the maxbcg cluster catalog is selected from imaging data contained in dr5 of the sdss .", "selection of galaxy clusters from this imaging data is done using the `` maxbcg '' algorithm .", "details of the algorithm are presented in @xcite , while the catalog and a description of its properties may be found in @xcite . in brief , the algorithm exploits two well - known features of rich galaxy clusters in addition to the tight spatial clustering of cluster galaxies .", "first , the bright end of the cluster luminosity function is dominated by red sequence galaxies with a small dispersion in color - magnitude space ( the e / s0 ridgeline ) .", "second , clusters contain a distinct brightest cluster galaxy ( bcg ) located near the center of the galaxy distribution .", "while some clusters lack an obvious central , dominant galaxy , every cluster does possess some red sequence galaxy brighter than any other .", "the algorithm measures independently the likelihood that a galaxy is spatially located in an overdensity of e / s0 ridgeline galaxies with similar @xmath35 and @xmath36 colors , and that it has the color and magnitude properties of a typical bcg .", "both likelihoods are evaluated for every sdss galaxy at a grid of redshifts .", "the redshift which maximizes the product of these likelihoods is then found for each galaxy .", "for e / s0 galaxies , this corresponding maximum likelihood redshift then provides a good estimate of the cluster redshift .", "once this list of cluster center likelihoods is assembled , these potential centers are ranked by decreasing maximum likelihood . the first cluster is seeded on the highest likelihood center , that cluster s bcg .", "galaxies projected within a scaled radius , @xmath37 , of this bcg and within @xmath38 in @xmath31 are eliminated from the list of potential centers .", "@xmath37 is the radius interior to which the mean density is 200 times the critical density ( @xmath39 ) as determined from sdss galaxy populations  @xcite .", "the process is repeated for the next most likely bcg on the list , given that it has not been eliminated by a higher likelihood bcg , and continues likewise until all potential centers have either been labeled as cluster bcgs or have been subsumed by higher likelihood centers .", "each cluster defined in this way has a center defined as the bcg location , an estimated redshift , and a richness , @xmath6 , given by the number of e / s0 ridgeline members falling within @xmath37 of the bcg and brighter than 0.4 @xmath40 .", "the final cluster catalog contains an array of measured properties , including photometric redshifts , richnesses , optical luminosities , and locations .", "the public maxbcg catalog contains a total of 13,823 clusters drawn from approximately 7500 square degrees of sky between redshifts of 0.1 and 0.3 , with a median redshift @xmath41 .", "the center for each cluster is defined as the location of the bcg identified by the algorithm .", "the richness of the cluster , @xmath6 , ranges between 10 and 188 in the public catalog , and in principle extends down to @xmath42 where the maxbcg selection function is less well - understood . for this study", "we include some clusters of lower richness , adding an additional 3532 clusters with @xmath43 .", "this slightly extended catalog allows us to use the same richness bins studied in the analysis of maxbcg galaxy dynamics  ( b07 ) and gravitational lensing  ( s07 , j07 ) .", "redshift estimates for the clusters , produced as part of the cluster finding process , have been shown by comparison to spectroscopic redshifts to be quite accurate , with @xmath11 @xcite .", "catalog completeness and purity have been studied in some detail in @xcite ; both are quite high .", "completeness is estimated to be @xmath44 for clusters with masses greater than @xmath45 , and purity is @xmath44 for clusters with richnesses @xmath46 .    the cluster population in this catalog has been used to derive constraints on cosmology @xcite using cluster counts .", "the relationship between the maxbcg richness @xmath6 and mass has been studied through galaxy dynamics  ( b07 ) and weak lensing ( s07 , j07 )", ". further work on galaxy populations , mass - to - light ratios , and improved richness estimates for these clusters is in progress .", "the _ rosat _ all - sky survey  ( rass , * ? ? ?", "* ) took place primarily during a six - month campaign in 1990 - 1991 to image the whole sky in soft x - rays ( 0.1 - 2.4 kev ) with the rosat position sensitive proportional counter  ( pspc , * ? ? ?", "the survey scanned the sky in great circles , with the largest net exposure time ( @xmath47 ) near the ecliptic poles .", "the typical field coincident with the maxbcg survey region , which does not overlap the northern ecliptic pole , has an effective exposure time of @xmath48 .", "the point spread function ( psf ) for these rass scans is very broad ( full - width - half - maximum of @xmath49 ) and is dominated by far off - axis photons due to the survey strategy .", "@xcite released reprocessed photon maps and exposure maps of the entire rass survey region .", "these photon maps provide the input for the analysis described in this paper .", "rass data has been used to create several catalogs of purely x - ray selected objects .", "the rosat bright source catalog  ( bsc , * ? ? ? * ) consists of 18,811 sources with a typical signal - to - noise @xmath50 .", "the position resolution is superior to the psf fwhm with 68% ( 90% ) of the sources within @xmath51 ( @xmath52 ) .", "in addition to the bsc , there is a companion faint source catalog  ( fsc , * ? ? ? * ) consisting of 105,924 sources with a typical signal - to - noise @xmath53 .", "most are unresolved .", "combined , these catalogs provide soft x - ray detected sources which can also be compared to the maxbcg clusters .    the rass photon data has also been used as an input for x - ray flux - limited cluster catalogs . the brightest cluster sample  ( bcs , * ? ? ?", "* ) is a flux - limited sample of the brightest 201 clusters in the northern hemisphere with fluxes @xmath54 .", "the northern rosat all - sky galaxy cluster survey  ( noras , * ? ? ?", "* ) is a catalog of 378 extended x - ray sources that have been confirmed to be clusters via optical follow - up . due to the broad rosat psf", ", this catalog is only @xmath55 complete at their stated flux limit , @xmath56 , which corresponds to a luminosity @xmath57 at our median redshift @xmath41 . @xcite", "have performed an initial comparison between noras clusters and maxbcg clusters and have found that the maxbcg detects @xmath58 of noras objects .", "this is consistent with estimates of the completeness from simulations  @xcite .", "the _ rosat_-eso flux limited x - ray galaxy cluster survey  ( reflex : * ? ? ?", "* ) is similar to noras in the southern sky , with the same flux limit .", "this catalog of 447 clusters is over 90% complete due to improvements in rass analysis , although only a small fraction of reflex overlaps the maxbcg survey area ."], ["the typical rass exposure time for maxbcg clusters , 400 s , is too short to allow significant detections for individual clusters .", "the large number of maxbcg clusters , however , allows us to make up for this .", "for example , there are 7986 clusters with richnesses 9 @xmath59 11 . for these objects ,", "the total rass exposure time is @xmath60 .", "such a large total exposure allows us to measure the mean x - ray emission from these clusters quite precisely .", "details of the stacking method are described later in this section , and outlined here .", "we begin by dividing the clusters into nine richness ( @xmath6 ) bins . to simplify comparison to other maxbcg analyses", ", we use the same richness bins used in measurements of the mean velocity dispersions  ( b07 ) .", "the number of clusters in each richness bin is shown in table  [ tab : ncluster ] .", "cccccc @xmath61 & 92.85 & 0.21 & 1727 & 55 & 55 + @xmath62 & 58.22 & 0.21 & 1469 & 146 & 140 + @xmath63 & 44.67 & 0.21 & 1317 & 207 & 201 + @xmath64 & 35.74 & 0.21 & 1201 & 356 & 339 + @xmath65 & 28.57 & 0.22 & 1102 & 665 & 633 + @xmath66 & 22.70 & 0.22 & 997 & 1128 & 1060 + @xmath67 & 18.91 & 0.23 & 941 & 1141 & 1099 + @xmath68 & 13.88 & 0.23 & 823 & 5651 &", "5405 + @xmath69 & 9.80 & 0.23 & 727 & 7986 & 7566 + random & & & 7986 & 7529 +    as we do not have x - ray centers for individual clusters , we treat the bcg selected by the maxbcg algorithm as the center of each cluster and stack on these centers .", "in @xcite it was shown that most maxbcg - selected bcgs ( @xmath70 ) agree well with the center of x - ray selected clusters , which has also been seen for other optically - selected catalogs  ( e.g. * ? ? ?", "possible biases introduced by this assumption are discussed in section  [ sec : centering ] .", "every source and background photon in the stacked analysis is scaled and weighted to the median redshift of the clusters in the catalog , @xmath71 . with these weighted photon maps", "we construct stacked images in section  [ sec : images ] , radial profiles in section  [ sec : profiles ] , and spectra and luminosities in section  [ sec : spectra ] .", "we obtain rass photon data and merged exposure maps from the archives available at the high energy astrophysics science archive research center ( heasarc ) .", "the rass data and exposure maps are distributed in arbitrarily constructed @xmath72 degree fields with significant overlap . nevertheless , since we utilize a large background annulus extending to @xmath73 from the central bcg , a significant fraction ( @xmath74 ) of the maxbcg clusters fall across a field boundary .", "we have therefore built a tool that extracts all rass photons in a given aperture , from multiple rass fields if necessary , counting only once those photons which appear multiple times in overlap regions .", "after photon retrieval , the appropriate merged exposure maps ( mex file ) are used to calculate the effective exposure time ( @xmath75 ) at the position of each detected photon .    before stacking", "we exclude from the list a subset of maxbcg clusters which might bias our x - ray measurements for unphysical reasons .", "no cut on galactic absorption is required , as the sdss observations are restricted to high galactic latitude .", "the typical equivalent galactic hydrogen column density is a very low : @xmath76  @xcite for the maxbcg clusters , with a maximum of @xmath77 .", "although the typical rass exposure time is @xmath78 , there are some fields with significantly less .", "this makes source and background estimation difficult , while not adding significantly to the signal . for this reason", ", we reject all clusters with less than 200  s mean exposure time .", "this removes 4% of the total number of clusters , but less than 1% of the net exposure time .", "an important possible source of contamination for stacked cluster measurements is the presence of a few extremely bright foreground sources .", "we would like to reject regions contaminated by these sources .", "rass images of maxbcg clusters , especially at higher redshift , are often unresolved , making it difficult to remove point sources using extent information , so we proceed as follows .", "we first identify those rosat bright source catalog ( bsc ) sources in the survey area with count rates higher than that expected for emission from any cluster at redshift beyond 0.1 .", "this corresponds to a count rate of @xmath79 .", "there are 179 bsc sources in the maxbcg survey area with rosat soft - band ( 0.1 - 0.5 kev ) or hard - band ( 0.5 - 2.1 kev ) count rates above this limit . of these 179 sources ,", "only 13 are clusters as identified in the _ rosat _ brightest cluster sample  ( bcs , * ? ? ?", "* ) . of these 13 clusters ,", "only one of these objects is associated with a maxbcg cluster ( abell 2142 , the richest and one of the nearest clusters in the catalog ) .", "the other high - flux clusters are at redshifts @xmath80 .", "further visual inspection has confirmed that abell 2142 is the only maxbcg cluster associated with one of these extremely bright sources . for the stacking analysis", ", we reject all maxbcg clusters ( except for abell 2142 ) within @xmath81 of any of these bright sources , due to possible contamination from non - cluster photons .", "this removes only 0.6% of the maxbcg catalog .", "the virgo cluster is another important foreground source .", "this nearby , x - ray bright cluster dominates the rass emission over a significant region of sky , thus making it difficult to estimate the x - ray emission from background clusters .", "we therefore reject all clusters that fall within @xmath82 degrees of m86 or m87 .", "this is a relatively small effect , removing only 0.1% of the maxbcg clusters .", "table  [ tab : ncluster ] shows the number of clusters in each richness bin before and after the rejections outlined in this section , as well as the mean richness for each bin .    to stack the x - ray measurements of the clusters , rass photons from each cluster", "must be scaled and appropriately weighted . the projected physical distance to the bcg", "is calculated for each photon , and this scaled distance is used in image construction and radial profile calculations . each source and background photon", "is then weighted by a factor of @xmath83 , where @xmath84 is the luminosity distance to the cluster and @xmath85 is the luminosity distance to the median cluster redshift @xmath71 . for background estimation", "we use a fixed annulus with inner ( outer ) radius of @xmath86 ( @xmath87 ) , similar to that used for the noras catalog for individual clusters  @xcite .", "this background annulus corresponds to a physical distance of @xmath88 at our minimum redshift of @xmath89 and a @xmath90 at our maximum redshift of @xmath91 .", "we confirm that our results are not sensitive to the exact choice of background annulus . to calculate the weighted count rate ( @xmath92 ) in a given bin ( image , radial , or spectral ) for a cluster at redshift @xmath31 , we follow a procedure similar to @xcite , where @xmath93 , where @xmath75 is the exposure time at the position of each detected photon .    as a simple test of our stacking procedure", ", we also extract rass photons from a set of positions selected at random from the region covered by the maxbcg catalog .", "each such random position is associated with a real cluster redshift drawn from the catalog in the @xmath69 bin .", "the random points are analyzed identically to the real cluster positions .", "the background - subtracted stacked images are shown in figure  [ fig : imageplots ] .", "each image has a projected radius of @xmath94 .", "the images contain photons from the _ rosat _ hard band ( channels 51 - 201 , 0.5 - 2.0 kev ) as this band has the highest signal - to - noise for cluster emission .", "after background subtraction , the counts were put in @xmath95 bins . for display , each image has been scaled with histogram equalization to show similar background noise levels , as there are two orders of magnitude more clusters in the poorest bin than the richest bin .", "the contours are drawn @xmath96 and @xmath97 above background level .", "highly significant x - ray emission is seen in each stack of clusters .", "by contrast , in a stack of 7,566 random points there is no significant excess , and no sign of emission centered in the stack .", "it is worth noting that the average count rate in the @xmath69 bin is only @xmath98 , corresponding to a flux @xmath99 .", "there are a total of @xmath100 in excess of background from @xmath101 clusters ; an average of only 1 source count per cluster .", "this illustrates the power of the stacking exercise to probe very low count rates , and allows us to use rass to measure very low @xmath5 cluster emission even at a moderate redshift of @xmath102 .", "figure  [ fig : profiles ] presents the background - subtracted surface - brightness profiles for the stacked x - ray images in each richness bin .", "the radial profiles have been summed in @xmath103 ( @xmath104 ) bins inside ( outside ) @xmath105 . for reference", ", the dotted lines represent the @xmath106 background level .", "significant cluster emission is seen in each richness bin . for the richest bins ,", "this extends out beyond @xmath105 .", "the x - ray surface brightness profiles for the different bins look remarkably similar , except for the normalization which increases strongly with richness and the signal - to - noise which decreases slowly with richness .", "we fit the surface brightness profiles out to @xmath107 with a standard @xmath108 model , @xmath109 .", "the best - fit parameters are presented in table  [ tab : betafits ] , and are overplotted with dashed lines in figure  [ fig : profiles ] .", "the @xmath108 model results in a good fit in all cases .", "cccc @xmath61 & @xmath110 & @xmath111 & 37.7/72 + @xmath62 & @xmath112 & @xmath113 & 43.2/72 + @xmath63 & @xmath114 & @xmath115 & 36.2/72 + @xmath64 & @xmath116 & @xmath111 & 75.1/72 + @xmath65 & @xmath117 & @xmath118 & 46.3/72 + @xmath66 & @xmath119 & @xmath120 & 88.1/72 + @xmath67 & @xmath121 & @xmath122 & 38.9/72 + @xmath68 & @xmath117 & @xmath118 & 36.2/72 + @xmath69 & @xmath123 & @xmath124 & 63.4/72 +    the core radii , @xmath125 , for these @xmath108 model fits are surprisingly large , ranging from @xmath126 to @xmath127 .", "these are significantly larger than those typically seen for x - ray clusters  ( @xmath128 , e.g. * ? ? ?", "the @xmath108 parameters are typical for x - ray clusters  ( e.g. * ? ? ?", "* ; * ? ? ?", "there is a slightly significant ( @xmath129 ) trend in @xmath108 with richness , such that the richer clusters tend to have slightly larger values of @xmath108 .", "this slight trend is in the same sense as has been seen in previous work correlating @xmath108 with cluster mass  @xcite .", "these @xmath108-model parameters are similar to those measured by dkm07 , in which @xmath108 was not seen to be strongly correlated with richness .", "these inconclusive results are caused by two effects ; the broad point spread function for rass objects , and the offset distribution between the bcg and the x - ray cluster emission .      to study the effect of the rass psf on the stacked profiles , we test our analysis on bright rosat point sources .", "these sources were selected from the wga catalog  @xcite , which contains 88,621 well measured point - like sources selected from rosat pspc pointed observations .", "we take @xmath130 moderately bright ( @xmath131 ) point sources that overlap the sdss dr5 mask .", "these bright sources are nearby galactic sources as well as agn and quasars at moderate and high redshift , and are also well detected in rass .", "each point source is randomly assigned a redshift drawn from the maxbcg cluster catalog redshift distribution .", "we note that a nearby unresolved point source is indistinguishable from a distant unresolved point source .", "the point sources are then stacked in exactly the manner described for clusters above .", "figure  [ fig : ptsrc ] shows the radial profile of the stacked point sources ( diamonds ) , which is essentially a measure of the rass psf with the radial scaling calculated for the maxbcg cluster redshifts . while this radial profile is much more sharply peaked than that of the stacked clusters , the broad rosat psf scatters significant x - ray emission to distances @xmath132", "( @xmath133 at @xmath134 ) from the center .", "this psf smearing implies an effective minimum on the @xmath108-model core radius , as well as creating a possible bias in the calculation of the @xmath108 parameter .", "we note that @xmath135 of the selected wga point sources are within @xmath136 of maxbcg clusters .", "excluding the point sources that are neighboring these possibly extended x - ray sources does not significantly alter this result .", "the stacked , projected surface brightness profile may also be affected by our choice of cluster centers .", "there are several ways in which the bcg chosen by the maxbcg algorithm might be offset from the x - ray emission from the cluster .", "dynamically active clusters and clusters in the process of merging can have a significant offset between the x - ray centroid and the bcg .", "there is also the possibility that the maxbcg algorithm chose an incorrect center ( this is addressed more fully in section  [ sec : centering ] ) .", "in addition , there may be additional x - ray point sources such as agn that are associated with the clusters .", "all of these effects result in an effective optical / x - ray offset distribution that may bias the radial profiles .", "we model the optical / x - ray offset distribution by matching the maxbcg catalog to known x - ray sources . for this exercise ,", "we use the fact that many maxbcg clusters ( @xmath137 ) are associated with individual detections in the rosat bright source and faint source catalogs .", "most of these bsc and fsc sources have not been previously recognized as associated with clusters , primarily because they are too faint or too distant to be seen as significantly extended in rass .", "we match the maxbcg clusters to the bsc and fsc catalogs , allowing multiple x - ray sources to match to each cluster .", "this ensures that we find all x - ray sources associated with each cluster , as well as all possible random matches .", "we similarly match an equal number of random locations drawn from the maxbcg survey region to the bsc and fsc catalogs .", "figure  [ fig : ptmatchhist ] shows the projected offset distribution from the maxbcg clusters to the bsc and fsc catalogs .", "the solid ( dashed ) line shows a histogram of maxbcg ( random ) offsets in @xmath103 bins .", "the dotted line shows the residual distribution after subtracting the random matches from the cluster matches .", "the subtracted histogram in figure  [ fig : ptmatchhist ] shows a large excess of x - ray sources associated with the optical cluster centers .", "there is a tight core in which the bcg is within @xmath138 of an x - ray source , as well as a long tail extending out to @xmath139 .", "the x - ray source excess at large radius is likely to be associated with the maxbcg clusters .", "these sources comprise a mix of merging clusters , clusters with poorly identified centers , and associated point sources such as agn .", "the subtracted histogram is used as a first - order empirical radial distribution of x - ray sources associated with maxbcg cluster centers .", "we now estimate the additional effect on the radial profile due to the radial offset distribution of x - ray sources associated with the maxbcg clusters . we take each wga point source from section  [ sec : rasspsf ] and", "randomly alter its position so that the distribution is offsets is identical to the empirical radial distribution .", "although we have only measured the offset distribution for the brightest @xmath140 of maxbcg clusters , we assume this is a good proxy for the offset distribution between the x - ray emission and bcgs of all the maxbcg clusters .", "these offset point sources are run through our stacking analysis .", "figure  [ fig : ptsrc ] shows the radial profile of the offset point sources after stacking ( circles ) .", "this radial profile is reasonably well fit with a @xmath108 model , although the fit is poor within @xmath141 . at large radii", "the radial profile is well fit with @xmath142 .", "thus , at large radii it is not possible to distinguish between true _ extended _ cluster emission and x - ray _ point _", "sources convolved with the observed maxbcg ", "x - ray source offset distribution .", "this calls into question the utility of the @xmath108 model parameterization for this exercise .", "the stacked x - ray profile is a convolution of the rosat psf , the centering distribution , and the true extended x - ray emission .", "operationally , it is not possible to separate the contributions from these three components .", "therefore , the @xmath108 parameters that are the result of a stacking exercise such as this one , where most of the individual x - ray clusters are not detected , should be used cautiously .", "we perform spectral fitting of the photons in each stack to calculate x - ray luminosities .", "the spectra were analyzed using xspec version 11.3.2  @xcite .", "spectral analysis on the stacked rosat data involves a variety of complications , some of which are discussed in dkm07 .", "we outline some key issues here .", "all rass observations integrate exposure time across the _ rosat _ field - of - view . as a result", ", we use the spectral response file pspcc_gain1_256.rmf from heasarc , suitable for off - axis observations .", "when combined with the appropriate ancillary response file ( arf ) , we obtain the vignetting corrected spectral response for the off - axis photons that comprise our spectra .", "however , the exposure times extracted from the merged exposure maps for rass are already corrected for vignetting , using the mean spectrum of the x - ray background in the pspc band . in practice", ", this means that our spectra , fit with the standard rmf+arf combination , are corrected for vignetting _", "twice _  once by using the exposure times obtained from the merged exposure map , and once in the calculation of the arf file . by comparing the on - axis response file rsp from heasarc with the combination of the rmf+arf we can approximate the extra vignetting correction .", "this correction is a function of energy , and depends ( weakly ) on the spectrum of the observed source . by simulating cluster spectra of various temperatures ,", "we calculate the typical correction factor averaged across the rosat hard band ( 0.5 - 2.0 kev ) which is dominated by cluster flux .", "this correction factor is @xmath143 . using this correction factor results in very good agreement between our spectral analysis and the reflex count - rate to flux conversion tables of @xcite .", "we estimate that applying this correction factor to all flux and luminosity values determined from stacked spectra adds an additional systematic error of @xmath144 .    for our analysis", ", we use the rmf file described above , combined with an arf output with the ftool pcarf .", "as noted in dkm07 , each arf file is essentially identical , because each stacked cluster from rass samples photons from the entire rosat field of view .", "spectral files are then grouped with a minimum of 50 counts per bin ( after background subtraction ) to ensure valid results using @xmath145 statistical analysis .", "fits performed with xspec were restricted to the 0.1 - 2.1 kev range .", "the uncertainties in spectral fit parameters are 90% confidence errors , obtained by allowing all fit parameters to vary simultaneously .", "we obtain cluster spectra by summing all weighted cluster photons in both fixed physical apertures and scaled apertures of the optically determined @xmath37 ( see section  [ sec : maxbcg ] ) .", "the fixed aperture of @xmath146 is chosen as a reasonable fiducial value because this provides good signal - to - noise , and because the radial profiles do not appear to change significantly with richness .", "the @xmath37 value for each bin was taken as the median @xmath37 of all the clusters in the richness bin ; these values are reported in table  [ tab : lxr200 ] .", "background spectra are stacked with the same weights as the source spectra , using the annuli defined in section  [ sec : extraction ] .", "as described previously , each photon is weighted to the median redshift @xmath147 .", "unfortunately , @xmath148-correction of individual photons is not possible , due to the fact that the detection channel of the incident photons is most strongly dominated by the spectral response of the rosat instrument .", "the stacking procedure will therefore tend to `` smear out '' the incident cluster spectra .", "simple simulations of x - ray spectra using xspec show that this does not create a large effect on our best - fit spectral values , as our redshift range of @xmath149 is not particularly broad .", "similarly , we do not make any corrections for possible redshift evolution in x - ray luminosity ; we will address this further in section  [ sec : redshift ] .", "ccccccc @xmath61 & 92.85 & 1.73 & @xmath150 & @xmath151 & @xmath152 & 47.2/76 + @xmath62 & 58.22 & 1.47 & @xmath153 & @xmath154 & @xmath155 & 61.5/77 + @xmath63 & 44.67 & 1.32 & @xmath156 & @xmath157 & @xmath158 & 87.4/65 + @xmath64 & 35.74 & 1.20 & @xmath159 & @xmath160 & @xmath161 & 100.6/95 + @xmath65 & 29.57 & 1.10 & @xmath162 & @xmath163 & @xmath164 & 62.9/95 + @xmath66 &", "22.70 & 0.99 & @xmath165 & @xmath166 & @xmath167 & 113.2/97 + @xmath67 & 18.91 & 0.94 & @xmath168 & @xmath166 & @xmath169 & 77.3/88 + @xmath68 & 13.88 & 0.82 & @xmath170 & @xmath171 & @xmath172 & 182.3/134 + @xmath69 & 9.80 & 0.73 & @xmath173 & @xmath174 & @xmath175 & 127.8/112 +    ccccccc @xmath61 & 92.85 & 1.73 & @xmath176 & @xmath177 & @xmath178 & 40.5/61 + @xmath62 & 58.22 & 1.47 & @xmath179 & @xmath180 & @xmath181 & 40.1/59 + @xmath63 & 44.67 & 1.32 & @xmath182 & @xmath183 & @xmath184 & 85.6/61 + @xmath64 & 35.74 & 1.20 & @xmath185 & @xmath186 & @xmath187 & 92.0/81 + @xmath65 & 28.57 & 1.10 & @xmath188 & @xmath189 & @xmath190 & 78.1/84 + @xmath66 & 22.70 & 0.99 & @xmath191 & @xmath192 & @xmath193 & 86.0/94 + @xmath67 & 18.91 & 0.94 & @xmath194 & @xmath195 & @xmath196 & 73.6/76 + @xmath68 & 13.88 & 0.82 & @xmath197 & @xmath171 & @xmath198 & 182.3/134 + @xmath69 & 9.80 & 0.73 & @xmath199 & @xmath200 & @xmath201 & 137.6/114 +    the spectra are fit with an absorbed thermal plasma model  @xcite .", "the metallicity is fixed at 0.3 solar  ( e.g. * ? ? ?", "* ) , and the redshift is set to the median scaled redshift of 0.2296 .", "the luminosities are calculated in the rest - frame 0.1 - 2.4  kev band at the median redshift .", "we prefer to calculate the 0.1 - 2.4  kev luminosity rather than bolometric luminosity ( @xmath202 ) , due to the large extrapolations required to obtain bolometric values , which strongly depend on spectral temperatures ( see below ) .", "the best - fit spectral parameters for the scaled @xmath37 apertures are shown in table  [ tab : lxr200 ] , and the parameters for the fixed @xmath146 apertures are shown in table  [ tab : lx750 ] .    in order to account properly for the variations in cluster luminosity in each richness bin", ", we use bootstrap resampling to estimate the luminosity errors . in each richness bin , we run 2000 trials by sampling the same number of clusters in that bin with replacement . to save considerable time with processing , we did not recreate the entire stacking procedure for each resampling .", "instead , we take the individual cluster counts , scaled to the median redshift .", "we confirm that the average scaled count rate is a good proxy for luminosity in the 0.1 - 2.4 kev band ( see section  [ sec : scatter ] for details on this calculation ) .", "the 68% ( @xmath106 ) confidence interval obtained from the bootstrap is added in quadrature with the @xmath5 errors obtained from the spectral fits .", "only the richest ( @xmath61 ) bin is dominated by the bootstrap error calculation , due to the large range in richness and @xmath5 in the bin .", "figure  [ fig : ltplot ] shows the mean @xmath0@xmath203 relation for our stacked clusters ( within the scaled @xmath37 aperture ) , compared to the @xcite @xmath5@xmath204 ( 0.1 - 2.4 kev ) relation .", "we see the cluster temperature increase with luminosity , but our observed @xmath0@xmath203 relation is noticeably steeper than that of @xcite .", "the stacked x - ray temperatures appear to underestimate the expected temperature , especially at @xmath0 larger than a few times @xmath205 .", "this discrepancy highlights the challenge of measuring cluster temperatures with rosat , as well as the challenges of measuring stacked cluster temperatures .", "first , rosat has sensitivity only to soft x - rays .", "when the break of the bremsstrahlung spectrum , determined by the temperature of the hot gas , is above @xmath206 , then the x - ray temperature becomes difficult to constrain .", "the _ rosat", "_ temperatures of hot clusters are generally underestimated ; a particular example is abell 1689 .", "this bright cluster has a nearly isothermal profile in _", "xmm / newton _ observations with a temperature of @xmath207 ( consistent with the asca and _ chandra _ values ) , but has a best - fit _ rosat_/pspc temperature of @xmath208  ( e.g. * ? ? ?", "second , when we stack many non - isothermal clusters with different temperatures , our isothermal raymond - smith spectrum is no longer appropriate .", "for example , @xcite have shown for _ chandra _ observations how the spectroscopically - weighted temperature can differ significantly from the emission - weighted temperature for single clusters , with the better measured colder gas dominating the spectral fit .", "when these effects are combined , interpreting the average best - fit spectral temperature from a stack of hundreds of clusters is non - trivial .", "most importantly for our task of measuring the mean @xmath0 , the rosat luminosity in the 0.1 - 2.4 kev band is virtually insensitive to the cluster temperature .", "if we fix the spectral temperature to the value predicted by the @xcite @xmath5@xmath204 relation , then @xmath0 changes by @xmath209 in each bin , with no systematic bias with richness .", "similarly , we can follow the prescription of @xcite to convert the mean 0.5 - 2.0 kev rass count rate to flux and luminosity .", "this method assumes the @xcite @xmath5@xmath204 relation , and requires iteratively calculating the flux and temperature .", "the @xmath0 values thus obtained are consistent with those from the spectral fitting to within @xmath210 .", "all this suggests that calculating 0.1 - 2.4  kev @xmath5 obtained from _", "rosat_which relies primarily on photon counting  is more robust than calculating @xmath204 and extrapolating to @xmath202 .", "although a small @xmath148-correction ( @xmath144 ) is required to extrapolate from observer frame 0.1 - 2.4  kev to rest - frame luminosity , this correction is not very sensitive to the spectral temperature at the moderate redshift of the maxbcg clusters ."], ["we stack the clusters and calculate the 0.1 - 2.4 kev x - ray luminosity as described in the previous section .", "figure  [ fig : lxn200plot ] shows the mean @xmath0@xmath1 relation , both for a fixed @xmath146 aperture and for the scaled @xmath37 apertures .", "we find a strong correlation between @xmath0 and @xmath1 with both the fixed and the scaled apertures . in the low richness bins ( @xmath211 ) , the scaled aperture @xmath37 is approximately equal to the fixed @xmath146 aperture , and @xmath0 is approximately the same for both choices of aperture . in the high richness bins ,", "the scaled aperture @xmath37 is significantly larger than @xmath146 , yielding a slightly larger @xmath0 for the scaled apertures .", "for this reason , the @xmath0@xmath1 relation is slightly steeper using scaled apertures .", "the best - fit mean relationships are :    @xmath212    @xmath213    power law fits are chosen to pivot around @xmath214 to approximately decouple errors in slope and normalization .", "dkm07 also found that the mean x - ray luminosity of the 2mass clusters scales with optical richness ( @xmath17 ) . a direct comparison between the dkm07 relation and our relation is challenging , due to the different luminosity and richness definitions we employ , but a first - order conversion is possible .", "first , we convert the dkm07 bolometric luminosities ( their table 2 ) to _ rosat _ 0.1 - 2.4 kev luminosities using table 5 of @xcite . ]", "re - fitting for the @xmath0@xmath17 relation , we find @xmath215.@xmath17 relation does not deviate from a power - law at any richness , unlike the @xmath216@xmath17 relation in figure 10 of dkm07 .", "] as the 2mass catalog is not yet public , we can not make a direct comparison of @xmath6 to @xmath17 for individual clusters .", "futhermore , most of the clusters in the 2mass catalog are at @xmath217 , while the maxbcg catalog covers @xmath218 . at this time ,", "the most we can say is that the slope of the @xmath0richness relation is roughly similar between our work and dkm07 .", "comparisons to x - ray selected catalogs ( see section  [ sec : xrayselect ] ) as well as prior studies  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ) lead us to expect significant scatter in the x - ray luminosity  richness relation for individual clusters . with a large scatter ,", "the mean statistic used to calculate the stacked @xmath0-@xmath1 relation may be significantly biased relative to the median relation . to understand this bias", ", we first assume that there is a log - normal conditional probability @xmath219 for the x - ray luminosity of a cluster at richness @xmath6 : @xmath220,\\ ] ] with constant intrinsic scatter @xmath221 , and mean log luminosity that follows : @xmath222 where @xmath223 is the log normalization of @xmath5 at @xmath214 , and @xmath224 is the slope of the @xmath5@xmath6 relation .", "note that @xmath225 for a log - normal distribution is equivalent to the median ( and geometric mean ) of the distribution . for the duration of this paper", ", we employ the notation @xmath226 .    at a given richness", ", we wish to measure the median x - ray luminosity @xmath227 , the peak of the underlying log - normal distribution", ". however , the stacking exercise we have undertaken is fundamentally a calculation of the _ arithmetic mean _ ( @xmath0 ) at a given @xmath1 . for a log - normal distribution with median @xmath227 and intrinsic scatter @xmath221 ,", "the arithmetic mean is @xmath228 .", "thus , the stacked normalization is an overestimate of the median of the underlying distribution by a factor of @xmath229 .", "if the scatter is large , the stacked ( mean ) normalization will be dominated by the most luminous clusters and will be biased high .", "for example , an @xmath230 scatter indicates a @xmath231 bias .    to constrain this scatter ,", "we begin with measurements of x - ray flux ( and hence @xmath5 ) at the locations of each cluster with @xmath232 .", "though many of these are low signal - to - noise detections , they can be used to measure scatter .", "the idl astronomy library tool linmix_err is used to fit @xmath5 as a function of @xmath6 with intrinsic scatter @xmath221  @xcite .", "this tool uses a bayesian approach to linear regression with errors in x and y and is well behaved even when the measurement errors dominate .", "it also handles non - detections and upper limits in y. monte carlo simulations show that the selected @xmath6 cut provides a large enough richness range to constrain the slope and scatter , and provides larger signal - to - noise than the entire cluster catalog , as over @xmath230 of the 955 clusters with @xmath233 are detected at at least the @xmath106 level .", "furthermore , the fits are not sensitive to the precise richness cutoff chosen .", "to calculate the @xmath5 for each of these often marginally - detected clusters , we use a procedure based on the count - rate to flux conversion method from the reflex survey  @xcite .", "first , we calculate the 0.5 - 2.0 kev ( rosat hard channel ) count rate in a @xmath146 aperture ; the fixed aperture was used because it results in better signal - to - noise than a larger scaled aperture ( where we only see significant signal in the stacked profiles ) .", "the local background is calculated in a @xmath234 annulus using the sector - rejection method of @xcite .    in reflex the temperature", "is estimated in an iterative fashion from the luminosity , using the @xmath5@xmath204 relation of @xcite . as most of the maxbcg clusters", "do not have a significant flux ( or luminosity ) measurement , we can not calculate the individual cluster temperatures in this way . instead", ", each cluster temperature is approximated by the stacked spectral temperature from the appropriate richness bin , as shown in table  [ tab : lx750 ] .", "this ensures that our stacked cluster luminosities and individual luminosity estimates are on approximately the same footing .", "perhaps more importantly , changing the temperature does not change the 0.1 - 2.4 kev luminosities significantly  ( e.g. * ? ? ?", "for example , for the clusters with @xmath235 detections , if we follow the reflex iterative recipe to convert count rate to @xmath5 , rather than fixing the temperatures at the stacked values , the individual @xmath5 values change by @xmath236 .", "this is simply another example of how measuring temperatures with rosat is challenging .", "the 0.5 - 2.0 kev count rate is then converted to 0.1 - 2.4 kev luminosity using table 2 from @xcite with the equivalent hydrogen column density at the position of the cluster  @xcite . after converting to luminosity", ", a @xmath148-correction is applied using equation 4 from @xcite , which is a good approximation of table 3 in @xcite.z\\}^{1/2}$ ] . ]", "the @xmath148-corrections are not very large , at most 10% .", "we compared our method of calculating @xmath5 for individual clusters to the @xmath5 values obtained in the noras catalog ( see section  [ sec : xrayselect ] for details on the x - ray catalog matching ) .", "for the matched clusters the values are all consistent within errors with @xmath237 scatter , and a systematic offset of @xmath238 . the primary difference in our calculations", "is that the noras fluxes were calculated in an aperture obtained via growth curve analysis ( gca ) designed to obtain the best signal - to - noise for each cluster , while we use a fixed physical aperture .", "therefore , our fixed temperature and aperture provide an unbiased estimate of the cluster flux and luminosity even without _ a priori _ knowledge of extended cluster emission .", "individual cluster @xmath5 and @xmath6 values are shown in figure [ fig : linmix ] .", "we use linmix_err to estimate the power - law slope , normalization , and intrinsic scatter of the underlying distribution of the @xmath5@xmath6 relationship for the 955 richest clusters with @xmath239 .", "the best fit relation is : @xmath240 with an intrinsic scatter @xmath241 .", "the individual @xmath5 values are calculated within a fixed @xmath146 aperture , and thus this is to be compared to equation  [ eqn : lxn_750 ] .", "we test the robustness of this result by both splitting the input data into multiple independently fit subsamples and changing the richness threshold slightly .", "we find that the constraint on the scatter is robust , and the error bar is accurate .", "possible systematic biases in the constraint on @xmath221 are discussed in section  [ sec : coolcore ] .", "the scatter - corrected @xmath227@xmath1 relation is shown in figure  [ fig : linmix ] .", "the luminosities for the individual @xmath106 detections are plotted as solid circles , and the @xmath106 upper limits are plotted as empty circles . the typical error bar for detections", "is shown in the lower - right corner .", "contours showing the @xmath242 contours on the best - fit median relation are shown in dark gray , and the dashed ( dotted ) lines show the @xmath243 ( @xmath244 ) scatter constraints .", "the median relation as constrained by linmix_err has been converted to the equivalent mean relation by multiplying the normalization by @xmath229 , and is shown in light gray .", "the mean relation agrees well with the individual stacked bins ( squares ) , which are about @xmath231 brighter than the underlying median values .", "we must emphasize that we are measuring the _ observed _ scatter in the @xmath5@xmath6 relation as is appropriate to our catalog and stacking method .", "this scatter comes about not only due to intrinsic @xmath5 variations between the clusters ( including different morphologies and merging clusters ) , but is also due to point sources , cooling flows , and the projection of multiple clusters along the line of sight . for the @xmath245 richest clusters ( @xmath246 ) used in this exercise , the chance of projection is very low .", "we do not have any way to remove point sources or bright cool cores from individual clusters in an unbiased way .", "operationally , it is not relevant to our measurement whether the scatter is due to intrinsic @xmath5 variations or due to point source and cool - core contamination . to constrain the observed scatter ,", "we need only assume that the underlying distribution is approximately log - normal , which is consistent with our observations .", "@xcite measure the observed scatter in the @xmath5@xmath247 relation , where @xmath247 is a richness measure from the amplitude of galaxy - cluster correlation function  @xcite .", "they use 290 optically - selected clusters , of which 40 have significant detections in _ rosat_/pspc observations . using a bayesian maximum - likelihood fitting technique", ", they find a significant correlation between @xmath5 ( bolometric ) and @xmath247 , with a power - law slope of @xmath248 and an intrinsic scatter of @xmath249 .", "it must be noted that they introduce a prior weighting , @xmath250 , which gives more weight to models with lower scatter .", "again , it is difficult to directly compare our measurement of the slope and scatter to the @xcite measurements , due to different richness quantities and selection functions .", "a large intrinsic scatter in the @xmath5@xmath6 relation can account for optically - selected clusters that appear `` underluminous '' in x - rays as compared to their optical richness  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "* ) . in particular , @xcite present a set of `` abell x - ray underluminous '' ( axu ) clusters with masses determined from velocity dispersion measurements , that are significantly dimmer in x - rays than expected . in figure", "[ fig : linmix ] the seven axu clusters that match to maxbcg ( @xmath2 , @xmath233 ) are denoted with open squares .", "these are all significantly dimmer than the mean @xmath5@xmath6 relation , but are fully consistent with the observed scatter .", "velocity dispersion measurements of maxbcg clusters  ( b07 ) and comparisons to simulations  @xcite have shown that there is considerable mass - mixing in a given @xmath6 bin .", "the low - luminosity clusters could be from lower - mass halos that are picked up as moderate richness in the cluster finder .", "however , @xcite argue that the axu clusters have low @xmath5 relative to their mass .", "this could be due to biases in masses calculated from velocity dispersions with small samples  ( e.g. * ? ? ?", "* ) , or to large intrinsic scatter in the @xmath5@xmath13 relation  ( e.g. * ? ? ?", "each of these possibilities ( or a combination of the two ) are consistent with the observations .", "we further explore the @xmath5@xmath13 relation of maxbcg clusters in a companion letter  @xcite .", "we have reason to believe that the scatter is not fixed as a function of richness .", "for example , b07 find a significantly larger scatter in the velocity dispersion - richness ( @xmath251@xmath1 ) relation at low richness compared to high richness", ". however , the quality of our data for this exercise is not adequate to constrain the change in the scatter as a function of richness .", "therefore , for the remainder of the paper we adopt this nominal value of @xmath241 to correct the stacked ( mean ) @xmath0 to obtain the underlying median @xmath227 values .", "this essentially requires us to multiply all @xmath0 values by a factor of @xmath252 .", "we can compare the @xmath227@xmath1 relation for optically - selected maxbcg clusters to that from x - ray selected clusters from the literature .", "for this exercise we chose to compare to the noras  @xcite and 400 square degree  ( 400d : * ? ? ? * ) catalogs , both constructed from rosat data , and thus measured in the same energy range as used in this work .", "the noras catalog is an x - ray flux limited cluster catalog constructed from the rass photon maps in the northern sky .", "although it is known to be only @xmath55 complete , it provides a large sample of bright x - ray clusters that overlap the maxbcg survey region .", "the 400d catalog is a serendipitous x - ray flux limited cluster catalog constructed from pointed rosat pspc observations .", "the 400d catalog covers fields sampled from the whole sky , and thus only @xmath249 of the catalog overlaps the maxbcg survey region .", "our intention in this exercise is to obtain a baseline comparison of the x - ray and optical richness properties of x - ray selected clusters to optically - selected clusters .", "a somewhat different comparison of noras and maxbcg , focused on testing the completeness of maxbcg , was performed in @xcite .    to obtain a list of clean matches between noras or 400d clusters and maxbcg clusters , we require that the x - ray position and bcg position be matched within @xmath253 , and the redshift difference to be less than 0.05 .", "this ensures both that the probability of a false match is @xmath254 and that we are not selecting , e.g. , merging clusters where the x - ray and optical catalogs might utilize different deblending schemes .", "there are 89 noras clusters that match maxbcg , with a median redshift of @xmath255 , and 53 400d clusters that match maxbcg , with a median redshift of @xmath256 .", "we convert all x - ray luminosities to 0.1 - 2.4  kev in the rest frame with our adopted cosmology .", "note that the @xmath5 values obtained in noras and 400d have been corrected for aperture effects .", "however , with the adopted @xmath108 parameters ( @xmath257 for noras , and @xmath258 for 400d ) these corrections produce at most a @xmath237 offset in @xmath5 compared to the luminosities obtained in our scaled @xmath37 apertures .", "figure  [ fig : catmatches ] shows @xmath5 vs. @xmath6 for the x - ray selected clusters that meet our matching criteria .", "the solid circles represent noras clusters , and the empty squares represent 400d clusters .", "the luminosities of the 400d clusters are typically less than that for noras due to the deeper flux limit of the pointed rosat observations relative to the rass survey .", "the dashed line shows the median @xmath227@xmath1 relation , and the dotted lines show the @xmath243 scatter constraints .", "it is not surprising to note that x - ray selection picks out primarily the x - ray brightest clusters at a given richness ; nearly all the noras clusters lie above the median maxbcg relation .", "the deeper 400d survey selects a sample more representative of the optically - selected clusters , though it is still biased high .", "comparison to figure  [ fig : linmix ] is instructive . at every richness , the typical x - ray emission from clusters is significantly below what you might expect from x - ray selected catalogs .", "while this is true for both catalogs , it is especially apparent for noras clusters .", "although the 400d survey has the sensitivity to detect the richest x - ray dim clusters ( as in the lower - right corner of figure  [ fig : linmix ] ) , these clusters are quite rare .", "the limited overlap between the 400d survey region and the maxbcg survey region means that it is very unlikely that the 400d survey would contain one of these specific maxbcg clusters ."], ["there are a variety of systematic effects which may bias the @xmath227@xmath1 relation . in this section", "we address six : photometric redshift uncertainty , cluster centering errors , richness variation with redshift , bcg luminosity , point source contamination , and cool core clusters .", "the maxbcg cluster catalog is based on photometric data , and thus we only have photometric redshifts ( `` photo-@xmath31s '' ) for each of the clusters . while the photo-@xmath31s are relatively precise ( @xmath11 ) , we must investigate how using photometric redshift estimates might bias our stacking results .", "if a cluster is actually closer than its photometric redshift would suggest , it will be over - weighted in the stacking analysis . when it is more distant than it seems , it will be underweighted .", "even if the photo-@xmath31 errors are perfectly symmetric , the weighting factor , @xmath83 , is not . as a result ,", "uncertainty in photo-@xmath31 can introduce a bias in the luminosities .", "this bias can become significant if the photo-@xmath31 errors are large .", "the net impact of this effect is to make the clusters appear more luminous than they truly are .", "there is also a small volume effect caused by photo-@xmath31 errors . at a fixed redshift @xmath31", "there is a larger physical volume at a larger distance @xmath259 than at a smaller distance @xmath260 .", "because of this , more clusters are available at high redshift to scatter low than at low redshift to scatter high .", "this is similar to malmquist bias .", "as above , this effect will cause us to overestimate the cluster luminosities .", "in addition , the photometric redshift is used to calculate the cluster extraction radius .", "when a cluster is closer than its photo-@xmath31 implies , then the extraction radius used is too small , slightly reducing its implied luminosity .", "the opposite is true when the cluster is farther than its measured photo-@xmath31 .", "as there is very little cluster flux at large radii , and as this bias is roughly symmetric , we do not expect this effect to strongly bias the stacked @xmath5 .", "we run a monte carlo simulation to estimate the combined photo-@xmath31 bias as a function of richness . around @xmath261 of the bcgs in the maxbcg catalog used in this analysis", "have spectroscopic redshifts from the sdss dr5 spectroscopic catalog  @xcite . with the spectroscopic subsample", ", we measure the photo-@xmath31 offset ( @xmath262 ) distribution in each richness bin and three redshift bins .", "a monte carlo is performed to estimate the overall bias in the measured luminosity by comparing the `` true '' luminosities convolved with the @xmath263 distribution to the `` observed '' luminosities at the measured photometric redshifts .", "we find that the overall bias is small ( @xmath264 at 99% confidence limit ) , with the observed luminosities slightly overestimating the true luminosities .", "this bias factor does not vary with richness .", "we neglect this correction factor due to its small magnitude .      in our stacking exercise , we chose the bcg as the center of the cluster .", "for many clusters with individual x - ray observations and a bright bcg , this is coincident with the x - ray center .", "however , there are several types of clusters in which the bcg is not consistent with the center of the x - ray emission .", "these include merging clusters with disturbed morphologies ; clusters without a dominant central galaxy for unambiguous detection by the maxbcg algorithm ; and a few ( @xmath265 ) previously identified strong cooling flow clusters .", "the central galaxies in these clusters host bright agn and exhibit strong star formation , moving their colors off the red sequence and excluding them from the maxbcg cluster finding algorithm .", "we do not expect centering biases to be a significant problem for our calculation of the mean x - ray luminosity for the following reasons .", "first , the x - ray luminosity is proportional to the electron density @xmath266 , and thus most of the luminosity comes from the central region . with our large apertures ( @xmath267 )", "as long as the core of a cluster is within the aperture , the photon counting exercise will include most of the cluster photons .", "second , we have reason to believe from simulations that most of the richer clusters are well centered  ( j07 ) .", "this is not to say that the cluster centering will not have any effect : as shown in section  [ sec : profiles ] , the @xmath108-model parameter is highly dependent on the centering distribution .    to confirm this , we perform some simple tests of de - centering the clusters . if we randomly offset all the cluster positions with a 2d gaussian with an rms of @xmath268  ( comparable to the centering distribution modeled in j07 ) , the mean @xmath0 values decrease by @xmath269 , and less than @xmath106", "this rms value was chosen because , as is seen in figure  [ fig : ptmatchhist ] , the vast majority of the x - ray cluster matches are within @xmath253 . therefore , we do not consider cluster miscentering to contribute a significant bias to our x - ray luminosity calculation .", "we investigate whether the observed @xmath0 at fixed richness changes with redshift .", "modest evolution in @xmath0 at fixed _ mass _ is expected .", "if the clusters are evolving in a self - similar manner , then we expect the higher redshift clusters to be more luminous than the lower redshift clusters , due to the fact that the universe was more dense at higher redshift .", "the expectation is that @xmath270 , where @xmath271 is the critical density of the universe at redshift @xmath31  @xcite .", "therefore , clusters at a given _ mass _ at a redshift @xmath272 should be @xmath273 brighter than similar clusters at a redshift of @xmath274 .", "to study possible variation of @xmath0 at fixed richness , we split the cluster sample into three redshift bins , containing the bottom 25% , middle 50% , and top 25% of clusters in our sample .", "the low redshift bin ranges from @xmath275 with a median redshift of @xmath276 ; the middle redshift bin ranges from @xmath277 with a median redshift of @xmath41 ; and the high redshift bin ranges from @xmath278 with a median of @xmath279 . to obtain increased signal - to - noise in these bins , we combine richness bins @xmath63 and @xmath62 ; @xmath65 and @xmath64 ; and @xmath280 and @xmath66 .", "these wider bins are the same that were used in the lensing analysis of s07 .", "figure  [ fig : splitzplot ] shows the stacked @xmath0 as a function of @xmath1 for the three different richness bins .", "the dashed line shows the mean relation ( equation  [ eqn : lxn_r200 ] ) for all clusters from section  [ sec : meanrelation ] .", "the high redshift clusters ( diamonds ) are significantly more luminous than the low redshift clusters ( circles ) .", "we parameterize the variation with a factor of @xmath281 , and fit all the redshift and richness bins simultaneously with a model of the form : @xmath282 where @xmath41 , the median redshift of the cluster catalog .", "this results in a good fit ( @xmath283 ) with the best - fit parameters : @xmath284 ; @xmath285 ; @xmath286 . note that @xmath287 and @xmath108 are consistent with the mean relation in equation  [ eqn : lxn_r200 ] .", "the redshift variation parameter , @xmath288 , is quite large , and shows that the high redshift bin ( @xmath279 ) is almost twice as bright as the low redshift bin ( @xmath276 ) .", "this is significantly in excess of the self - similar prediction .", "the challenge is to determine the origin of this redshift dependent shift in @xmath0 at fixed richness .", "the photo-@xmath31 errors described in section  [ sec : photoz ] may account for some of this effect .", "although the absolute photo-@xmath31 errors are better than @xmath289 at all redshifts , the _ relative _ photo-@xmath31 errors are larger at low redshift than at high .", "therefore , the impact of photo-@xmath31 bias is larger at low redshift than at high redshift .", "we repeated the monte carlo simulation described in section  [ sec : photoz ] to measure the photo-@xmath31 bias from each redshift bin .", "the additional redshift bias is @xmath290 ( 99% confidence limit ) in the low redshift bin , and @xmath291 ( 99% confidence limit ) in the high redshift bin .", "the photo-@xmath31 bias therefore remains smaller than the observed variation in @xmath5 , although the effect is in the same sense as the variation we observe , and is a contributing factor .", "a more likely possibility is redshift dependent variation of @xmath6 at fixed mass .", "if our high redshift clusters have systematically smaller @xmath6 at fixed mass , this would have the effect of shifting the high redshift points in figure  [ fig : splitzplot ] to the left . in order to constrain this bias ,", "we must first factor out the effect of the self - similar evolution of @xmath0 at fixed mass , using the expected redshift dependence @xmath292 as described above .", "we then re - fit the data to equation  [ eqn : zgamma ] to calculate the excess variation in the @xmath0@xmath1 relation that may be attributed to variation of @xmath6 at fixed mass .", "this results in a best - fit @xmath293 , implying a fractional decrease in @xmath6 of @xmath294@xmath295 from our lowest redshift bin ( @xmath296 ) to our highest redshift bin ( @xmath297 ) .", "this is consistent with the redshift dependent variation in the velocity dispersion ", "optical richness relation measured by b07 .", "however , we must note that redshift dependence in the observed lensing shear , @xmath298 , is significantly smaller  ( s07 ) .", "unfortunately , none of these approaches is yet able to confidently determine the nature of the observed variation .", "for example , this effect could be caused by incorrect handling of the 0.4  @xmath18 and color cuts that determine which galaxies are included in the richness estimate .", "it might also be `` true '' evolution , such that clusters have fewer red - sequence galaxies brighter than 0.4 @xmath18 at fixed mass at higher redshifts .", "further work , including improving the richness estimates , is in progress to constrain the nature of this evolution .", "we now investigate the effect bcg @xmath30-band luminosity ( @xmath299 , see @xcite for details ) has on the mean x - ray luminosity @xmath0 .", "simulations and semi - analytical modeling predict that dark matter halos formed at early times have brighter bcgs and lower richness than those which form late  @xcite .", "thus , we expect that the x - ray luminosity should be correlated with @xmath299 at fixed richness .", "furthermore , optical and x - ray observations of individual nearby clusters show that more luminous bcgs are correlated with more massive halos with higher x - ray temperature and luminosity  ( e.g. * ? ? ?", "this trend has already been seen in the mean velocity dispersion analysis of maxbcg clusters , as the clusters with more luminous bcgs had significantly larger velocity dispersions than clusters with similar richness  ( b07 ) .    in an exercise similar to that performed in the previous section , we split each wide richness bin into three bins of @xmath299 .", "as @xmath299 , unlike redshift , is correlated with @xmath6 , we are unable to use the same @xmath299 split for each @xmath6 bin .", "therefore , we sort the clusters in each richness bin by @xmath299 , and split the sample into the top 25% , middle 50% , and bottom 25% , and then restack .", "the results are shown in figure  [ fig : splitbcgplot ] . it is readily apparent that the clusters with the most luminous bcg in a given richness bin ( diamonds ) are significantly more luminous in x - rays than other clusters in the bin .", "this effect is most dramatic at lower richnesses .", "this is expected , as the bcg is more dominant in low richness clusters than high richness clusters  ( e.g. * ? ? ?", "it is also notable that there does not appear to be a significant difference between the @xmath0 for the low @xmath299 and middle @xmath299 bins .", "we parameterize the effect of @xmath299 on @xmath0 using a similar procedure as performed for the redshift variation .", "the richness and @xmath299 bins are fit simultaneously with a model of the form : @xmath300 this results in an adequate fit ( @xmath301 ) with the best - fit parameters : @xmath302 ; @xmath303 ; @xmath304 .", "although the normalization is the same as in equation  [ eqn : lxn_r200 ] , the slope is significantly more shallow .", "this is caused by the correlation between @xmath299 and @xmath6 .", "after taking this into account , the mean relation shown here is fully consistent with that calculated without splitting the bins by @xmath299 .", "the reason the @xmath145 of the fit is relatively high is because it does not appear that @xmath0 is smooth function of @xmath305 : the effect of @xmath299 is not symmetric , and only tends to boost the x - ray luminosity of clusters with relatively bright bcgs .", "it is not entirely clear whether this effect is due to an imperfect cluster finder , our crude richness definition , or due to real cluster physics .", "clusters with brighter , more dominant bcgs are more likely to be correctly centered and thus might be more luminous in x - rays .", "however , as we showed in section  [ sec : centering ] , the effect of decentered clusters on @xmath0 is much smaller than the effect we see due to @xmath299 .", "meanwhile , our richness definition , @xmath6 , is a count of red - sequence galaxies brighter than @xmath306 .", "it is thus not surprising that a poor cluster with an exceptionally bright bcg might be more massive and more x - ray luminous than a cluster at a similar richness with a more typical bcg .", "finally , it is possible that these clusters are associated with halos that formed early , allowing many of their member galaxies to merge into a very large bcg  @xcite . deeper targeted x - ray observations of specific clusters will be required to determine if we can use @xmath299 and other optical properties to constrain the age of its dark matter halo .      when calculating the projected x - ray luminosity from the hot intracluster medium , there is the possibility of contamination due to x - ray point sources .", "some point sources , such as cluster agn , are associated with the cluster , but in general their luminosities are not tightly correlated with that of the icm .", "other point sources that are chance coincidences , such as foreground stars and background quasars , might also boost the apparent luminosity .", "the problem of point source contamination is exacerbated by the broad psf of the rass survey .", "due to the large psf , it is difficult to accurately excise point sources from the higher redshift clusters , even when their positions are known . for the stacking exercise ,", "foreground and background sources are not a significant problem . assuming the positions of these point sources are uncorrelated with the cluster positions , the contribution from these sources is subtracted with the background estimation , as is demonstrated by our stack of random points in figure  [ fig : imageplots ] . at the same time , these point sources , both those associated with the clusters and those that are chance projections , may increase the observed scatter . in general", ", we note that @xmath5 values calculated from rass are not corrected for point source contamination  ( e.g. * ? ? ?", "* ; * ? ? ?", "we wish to perform a first - order check on possible contamination from cluster agn , the dominating point sources that might bias the observed @xmath0 .", "@xcite have recently performed a detailed survey of the distribution of agn in galaxy clusters for eight moderate redshift clusters , comparable to the redshift range of the maxbcg catalog .", "they find that moderately bright agn ( @xmath307 ) associated with bright cluster members ( @xmath308 ) make up around @xmath309 of the total cluster member population .", "this implies that many of the richest maxbcg clusters contain at least one moderately bright agn .", "these clusters have an extremely bright icm ( @xmath310 ) , and thus the fraction of the luminosity from cluster agn will , on average , be very small .", "if we extrapolate the same agn fraction to our poorest clusters ( @xmath311 , perhaps one in ten will host a moderately bright agn . although these agn might have a luminosity comparable to the icm for these @xmath312 of poor clusters , this small fraction should not strongly bias the stacked average .    as a simple check for contamination in the x - ray signal from cluster agn", ", we match maxbcg cluster member galaxies to radio sources from the first survey  @xcite with a match radius of @xmath313 . by matching random galaxies from sdss dr5 to the first sources", ", we estimate the purity of the matches to be @xmath314 .", "the first survey covers approximately the same footprint as the sdss , down to a typical flux limit of @xmath315 at 1.4 ghz .", "using 1.4 ghz radio detections as a proxy for agn activity has the advantage that the radio spectrum is often flat , and thus the 1.4 ghz flux is not a strong function of redshift .", "the disadvantages are that only 10% of x - ray bright quasars and agn are radio loud , and that there is a large scatter between x - ray and radio luminosities of these objects . for example", ", @xcite noted that for the same small set of clusters none of the x - ray selected agn are radio loud  @xcite .", "although we have an incomplete selection function , we can still constrain the contamination from cluster agn . if the mean @xmath0 is significantly boosted by cluster agn", ", then we would expect the clusters which match radio sources to be relatively bright .", "this effect would be greatest at low richness and low @xmath0 when the icm is not as hot or bright .", "for each richness bin , we matched the maxbcg cluster members within @xmath146 of the bcg to sources in the first survey .", "table  [ tab : radiomatch ] shows the number of clusters in each bin that have members matched with the first catalog .", "more detailed work is in progress in cross - correlating the maxbcg catalog with radio sources to explore cluster agn feedback and related issues ( see also @xcite ) .", "ccc @xmath61 & 45 & 82% + @xmath62 & 91 & 65% + @xmath63 & 125 & 62% + @xmath64 & 183 & 54% + @xmath65 & 284 & 45% + @xmath66 & 420 & 37% + @xmath67 & 332 & 30% + @xmath68 & 1330 & 25% + @xmath69 & 1341 & 18% +    we wish to compare how the radio selected subset of clusters compares to a randomly selected subset of clusters of similar richness . in each richness bin , we ran 10000 trials by sampling ( with replacement ) from the entire set of clusters in the bin , sampling the same number of clusters that match radio sources . as a quick estimate of the mean luminosity , rather than performing the full stacking analysis and spectral fits , we calculated the stacked and weighted , background - subtracted hard channel ( 0.5 - 2.0 kev ) counts in a fixed @xmath146 aperture .", "the resulting count rate was converted to @xmath0 using the method described in section  [ sec : scatter ] .", "we can then compare the mean @xmath0 of the clusters which match radio sources to the typical values of @xmath0 that we would expect when drawing that same number of clusters from the entire sample .", "the results are shown in figure  [ fig : radiomatch ] . in each panel", "the histogram shows the distribution of @xmath0 values obtained from the resampling , and the vertical dashed line marks the mean @xmath0 of the radio - matched clusters .", "we do not see any significant bias as a function of richness .", "the only bin in which the radio matches appear to be outliers is the @xmath316 richness bin .", "this is due to random chance , in that the clusters in this richness bin with the highest @xmath5 also happen to match radio sources ; although there is no obvious indication in the rass data that there is point source contamination on these clusters , further follow - up with _ chandra _ or _ xmm / newton _ would be required to clarify this .", "there is much evidence that cool core clusters increase the scatter in x - ray cluster properties . by excluding cluster cores in high resolution imaging ,", "many x - ray parameters become more tightly correlated  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ; * ? ? ?", "the broad rosat psf , combined with the fact that most of our clusters do not have individual x - ray detections , makes it impossible to exclude cluster cores in our stacking analysis . however , a few of the brightest clusters ( @xmath317 ) at moderate richness ( @xmath318 ) are well known cool - core clusters .", "we thus investigate what effect cool core clusters have on our estimate of the median @xmath227 and scatter @xmath221 in the @xmath227@xmath1 relation .", "there does not exist an unbiased volume - limited catalog of cool core x - ray clusters which can be fairly compared to the maxbcg catalog .", "@xcite used rosat pointed observations of an x - ray flux limited catalog  @xcite to estimate the central cooling time of 55 nearby clusters .", "two of these clusters are in the maxbcg catalog ( a1689 , a2244 ) and exhibit characteristics of a cool core ( @xmath319 ) .", "@xcite used _ chandra _ observations of the higher redshift ( @xmath320 ) bcs clusters to systematically estimate the central cooling time of bright x - ray clusters .", "we identify 7 maxbcg clusters ( a750 , a963 , a1835 , rxj 2129.6 + 0005 , z2701 , z3146 , z7160 ) from this sample with cooling times @xmath319 which we mark as cool core clusters .", "finally , we identify clg  j1504 - 0248  @xcite as maxbcg cluster with a cool core .", "it should be noted that most of these clusters have a moderate richness ( @xmath321 ) and are in the brightest tail of the @xmath5 distribution for their richness .", "furthermore , the central galaxies of these clusters tend to show strong @xmath322 emission in sdss spectroscopy .", "in fact , for a few of these clusters the true bcg is not in the list of cluster members because the strong @xmath322 emission changes the apparent galaxy color , making it inconsistent with the red - sequence used in the cluster finding algorithm .    after excluding these known cool core clusters ( which might bias the @xmath227@xmath6 relation ) we recalculate the stacked mean relationship as well as the median relationship with scatter . when calculating the median relationship with scatter , as in section  [ sec : scatter ] , the normalization and slope shift by @xmath323 , as these values are not strongly affected by outliers .", "however , the intrinsic observed scatter decreases to @xmath324 , which is a @xmath325 shift .", "it is not surprising that @xmath221 decreases : we are deliberately removing the brightest clusters .", "meanwhile , the stacked relation is also slightly affected by these bright outliers , and after cutting the known cool core clusters the normalization of the stacked relation decreases by @xmath326 . after following the previously described prescription for converting the stacked mean relation to a scatter - corrected median relation", ", we find that removing the known cool core clusters decreases the normalization by @xmath209 .", "therefore , it does not appear that the known , very x - ray luminous , cool core clusters , are significantly biasing our relation .", "we reiterate that the intrinsic scatter we measure is the _ observed _ scatter that takes into account _ all _ x - ray observations of the maxbcg clusters , regardless of x - ray morphology of the clusters .", "thus , it is likely that some portion of this observed scatter is caused by unidentified cool core clusters in the sample ."], ["b07 have measured the median velocity dispersion ( @xmath328 ) as a function of @xmath1 for the maxbcg catalog .", "the richness bins used to measure the stacked velocity dispersions , as well as the scatter in the @xmath12@xmath6 relation , are the same as those used in this work . however , in order to measure velocity dispersions , only clusters that have the bcg and at least one additional member galaxy in the sdss dr5 spectroscopic subsample  @xcite were used .", "the spectroscopic subsample preferentially selects bright / nearby galaxies , and thus the selection of clusters used in b07 ( @xmath329 , with @xmath330 ) is slightly different than that for the full maxbcg catalog ( @xmath331 , with @xmath332 ) . as discussed in section  [ sec : redshift ] , the lower redshift clusters have a smaller @xmath0 than the higher redshift clusters .", "we therefore re - run our stacking procedure on exactly those clusters used in b07 , and use the known spectroscopic redshifts of the bcgs rather than the photometric redshifts .", "this ensures that , for this exercise at least , we are not affected by any possible photometric redshift biases , which otherwise could be significant for the nearest clusters ( @xmath333 ) where the relative photo-@xmath31 error ( @xmath334 ) is quite large . table  [ tab : lxsig ] shows the richness bins used , as well as the mean @xmath0 ( this work ) and median @xmath327 ( from b07 ) for the nine richness bins .", "the slope of the @xmath0@xmath1 relation is nearly identical for the spectroscopic subsample as reported in section  [ sec : meanrelation ] , but the normalization is lower by @xmath312 , due to the different redshift selection .", "ccccc @xmath61 & 83.9 & 1.69 & @xmath335 & @xmath336 + @xmath62 & 58.4 & 1.45 & @xmath337 & @xmath338 + @xmath63 & 44.7 & 1.32 & @xmath339 & @xmath338 + @xmath64 & 35.9 & 1.22 & @xmath340 & @xmath341 + @xmath65 & 28.6 & 1.10 & @xmath342 & @xmath343 + @xmath66 & 22.7 & 1.02 & @xmath344 & @xmath345 + @xmath67 & 19.0 & 0.94 & @xmath346 & @xmath347 + @xmath68 & 14.0 & 0.82 & @xmath348 & @xmath349 + @xmath69 & 9.9 & 0.73 & @xmath350 & @xmath351 +    we take the mean @xmath0 values and corrected them for scatter as described in section  [ sec : scatter ] to obtain @xmath227 .", "figure  [ fig : lxsig ] shows the median @xmath227@xmath327 relation for maxbcg clusters with richness @xmath352 . each data point ( solid circles ) is obtained from the same richness bin and same cluster selection .", "the dark gray band shows the best - fit relation ( @xmath353 ) from clusters in the spectroscopic subsample of the maxbcg catalog , with the following functional form : @xmath354 the scaling relation predicted by self - similar cluster evolution is that @xmath355  ( e.g. * ? ? ?", "* ; * ? ? ?", "* ) , which is consistent with our observations .", "we note that if we could convert our 0.1 - 2.4  kev luminosities to @xmath202 , this would steepen our relation , as the clusters with higher velocity dispersions also have a higher temperature and larger @xmath356 ratio .", "as mentioned previously , such a conversion is not practical for the stacking exercise using rosat data .", "the slope of our 0.1 - 2.4 kev @xmath227@xmath327 relation is consistent with that measured for reflex clusters  @xcite .", "we compare this median relation to that obtained previously for individual x - ray selected clusters .", "@xcite calculated the velocity dispersions of reflex clusters by cross - correlating the reflex catalog with the two - degree field galaxy redshift survey  ( 2df ; * ? ? ?", "figure  [ fig : lxsig ] shows 39 individual reflex clusters ( empty circles ) from hilton , et al .", "( 2007 , private communication ) after converting the velocity dispersions to the cluster rest frame . in the upper - left corner", "the empty circle shows the typical errors in 0.1 - 2.4 kev @xmath5 and @xmath12 for the reflex/2df clusters .", "@xcite found the slope of the x - ray selected @xmath5@xmath12 relation to be @xmath357 , consistent with our observations .", "however , the individual x - ray selected clusters tend to have a slightly higher ( @xmath326 ) velocity dispersion at a given x - ray luminosity compared to the stacked maxbcg clusters .", "this is possibly caused by the difficulty in measuring velocity dispersions for individual clusters .", "contamination from non - virialized galaxies and interlopers will more likely result in an overestimate of the velocity dispersion rather than an underestimate ."], ["by stacking observations from the _ rosat _ all - sky survey , we measure the mean x - ray luminosity as a function of richness for optically - selected clusters from the maxbcg catalog . with a large number of clusters in each bin , the stacking exercise has the power to probe to much lower flux limits than are possible for x - ray selected surveys .", "thus , a highly pure and complete volume - limited catalog with clusters selected by their optical properties can be used to measure the mean @xmath0 .", "we find that @xmath0 scales with optical richness , @xmath6 , with a simple power - law form over two orders of magnitude in @xmath0 .", "our results are similar to those obtained in dkm07 by stacking rass observations of clusters identified in 2mass .", "however , we find that stacked temperature measurements from _ rosat _ are significantly biased for higher temperature clusters .", "therefore , we restrict our analysis to luminosities calculated in the 0.1 - 2.4  kev band rather than extrapolating to calculate bolometric luminosities . furthermore , we have shown that the @xmath108-model fits to the stacked radial profiles of moderate redshift clusters are dominated by a combination of the broad _ rosat _ psf and the offset distribution between maxbcg clusters and correlated x - ray sources . on the other hand , calculating x - ray luminosity in the 0.1 - 2.4 kev band , which relies primarily on counting photons , is not strongly biased by spectral fits or profile fits . as @xmath5 scales with the square of the density , as long as the core of a cluster is within the stacking aperture , we can obtain a high reliability estimate of the mean cluster luminosity .    by measuring the x - ray flux at the positions of individual maxbcg clusters , we are able to constrain the observed scatter in the @xmath5@xmath6 relation . assuming a log - normal distribution of @xmath5 as a function of @xmath6 , we find @xmath241 .", "this scatter is quite large , and thus it is necessary to correct for the scatter when calculating the median of the underlying distribution @xmath227 .", "the richness measure we have used in this work , @xmath6 , is simply the count of red - sequence galaxies within a scaled @xmath37 aperture , but nevertheless is a good proxy for @xmath5 on average .", "it is likely that improved richness measurements can be made . from the evidence presented here , it seems likely that these new richness estimators will include information about bcg luminosity , and will more carefully control variations in cluster membership with redshift .", "the large scatter in the @xmath5richness relation does have some significant effects on cluster selection .", "first , x - ray selection with a high flux limit , as with the noras catalog , tends to pick out the clusters with the highest @xmath5 at a given richness .", "second , the large scatter in the @xmath5@xmath6 relation means that there are a significant number of optically selected clusters that are seemingly `` underluminous '' .", "optically selected clusters will often appear relatively dim compared to x - ray selected clusters at a similar richness .", "the goal of cluster selection and identification is to find a low scatter proxy for halo mass .", "both x - ray and optical selection techniques suffer from different limitations .", "x - ray flux limited surveys can only find the brightest x - ray clusters , and are thus limited to a combination of the brightest and nearest clusters . furthermore , there may be significant scatter between x - ray luminosity and halo mass  @xcite .", "optical surveys such as maxbcg may be volume limited out to moderate redshift , but optical richness also has a large scatter with respect to x - ray luminosity and mass  ( b07 ) .", "the development of more precise mass proxies , including @xmath204 , the compton y parameter @xcite , @xmath358  @xcite , and potentially more sophisticated optical richness estimates may allow selection of cluster catalogs more closely approximating the mass - limited catalog we aspire to .    by combining our present analysis with the velocity dispersion measurements of b07", ", we have measured the median @xmath227@xmath327 relation for the maxbcg clusters .", "the relation has a slope of @xmath359 , consistent with the prediction of self - similar cluster evolution .", "previous measurements of the @xmath5@xmath12 relation have been complicated by both selection effects and a large scatter , as the derived slope depends strongly on the fitting technique used . by stacking clusters from a volume - limited catalog and correcting for scatter , both @xmath227 and @xmath327", "are much better constrained , allowing a robust calculation of the relation . in our analysis", ", we do not see any evidence of a break in the @xmath227@xmath327 relation at the poor cluster ( group ) scale , as has been hinted at previously  ( e.g. * ? ? ?", ". this might be due to the fact that the maxbcg cluster finder works with red - sequence clusters , and does not find groups with a large blue galaxy fraction .", "we can also compare @xmath0 as determined from this method to weak lensing masses determined from the same maxbcg cluster catalog  ( s07 , j07 ) .", "determining the @xmath5@xmath13 relation is not only important for understanding cluster physics , but for calibrating the selection function of x - ray clusters used as cosmological probes  ( e.g. * ? ? ?", "we have made this measurement in a companion letter  @xcite .", "in addition , the comparison of the x - ray luminosities , weak lensing masses , and velocity dispersions can be used in conjunction with the number function of maxbcg clusters to constrain cosmological parameters  ( e.g. * ? ? ? * ) .", "this work is ongoing .", "refined measurements of the scatter , point source contamination , and confirmation of catalog purity will require deeper pointed x - ray observations .", "we are currently investigating serendipitous maxbcg cluster observations by deeper pointed _ rosat_/pspc observations , as well as _", "xmm / newton _ and _ chandra _  ( e.g. * ? ? ?", "unfortunately , most of the serendipitous cluster observations are of relatively poor clusters , which are much more numerous .", "a targeted campaign of a representative sample of maxbcg clusters that have not previously been known to be x - ray bright will be essential to both test our measurement of the scatter in @xmath5@xmath6 and to estimate the fraction of cluster flux that is contaminated by point sources .", "this will be essential to enable future inexpensive deep optical cluster surveys such as des  @xcite to use optical properties of clusters to estimate which clusters would be most useful to follow - up with targeted x - ray observations .", "e. rykoff and t. mckay are pleased to acknowledge financial support from nsf ast-0206277 and ast-0407061 , and the hospitality of the michigan center for theoretical physics .", "mrb acknowledges the support of the michigan space grant consortium .", "aee acknowledges support from nsf ast-0708150 .", "we also wish to thank the anonymous referee for many helpful comments .", "funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .", "department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .", "the sdss web site is http://www.sdss.org/.    the sdss is managed by the astrophysical research consortium for the participating institutions .", "the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .                                                                                                                                                                , e. et  al .", "1987 , in presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol .", "733 , soft x - ray optics and technology ; proceedings of the meeting , berlin , federal republic of germany , dec .", "8 - 11 , 1986 , bellingham , wa , society of photo - optical instrumentation engineers , volume 733 , 1987 , p. 519 . ,", "koch & g.  schmahl , 519+                                          , w. , boller , t. , englhauser , j. , freyberg , m. , & supper , r. 2001 , in astronomical society of the pacific conference series , vol .", "225 , virtual observatories of the future , ed .", "r.  j. brunner , s.  g. djorgovski , & a.  s. szalay , 234+"]]}
