1: \documentclass[apj]{emulateapj}
2: \usepackage{amsmath}
3: \usepackage{tmsfnts}
4:
5: \def\kmsmpc {km$\;$s$^{-1}\,$Mpc$^{-1}$}
6: \def\wmap {\emph{WMAP}}
7: \def\chandra {\emph{Chandra}}
8:
9: \begin{document}
10:
11:
12: \submitted{Submitted to ApJ Letters, 2007 May 23}
13:
14: \lefthead{HELIUM ABUNDANCE IN GALAXY CLUSTERS}
15: \righthead{MARKEVITCH}
16:
17: \title{HELIUM ABUNDANCE IN GALAXY CLUSTERS AND SUNYAEV-ZELDOVICH EFFECT}
18:
19: \author{M.~Markevitch\altaffilmark{1}}
20:
21: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,
22: Cambridge, MA 02138; maxim@head.cfa.harvard.edu}
23:
24: \altaffiltext{1}{Also Space Research Institute, Moscow, Russia}
25:
26: \setcounter{footnote}{1}
27:
28: \begin{abstract}
29:
30: It has long been suggested that helium nuclei in the intracluster plasma can
31: sediment in the cluster gravitational potential well. Some theoretical
32: estimates for the cores of relaxed clusters predict an excess of helium
33: abundance by up to a factor of a few over its primordial value. The
34: intracluster helium abundance cannot be measured directly. This presents a
35: significant source of uncertainty for cosmological tests based on the X-ray
36: derived cluster quantities, such as the gas mass, total mass, and gas mass
37: fraction, all of which depend on the assumed helium abundance. We point out
38: that cluster distances derived by combining the Sunyaev-Zeldovich (SZ) and
39: X-ray data also depend on the helium abundance. This dependence can be used
40: to {\em measure}\/ the abundance, provided the distance is known
41: independently. For example, if one adopts the \wmap\ $H_0$ value, then the
42: recent $H_0$ measurement by Bonamente and collaborators, derived from SZ
43: data on 38 clusters assuming a primordial helium abundance, corresponds to
44: an abundance excess by a factor of $1.9\pm0.8$ within $r\sim 1$ Mpc (using
45: only their statistical errors). This shows that interesting accuracy is
46: within reach. We also briefly discuss how the SZ and X-ray cluster data can
47: be combined to resolve the helium abundance dependence for the $d_a(z)$
48: cosmological test.
49:
50: \end{abstract}
51:
52: \keywords{cosmic microwave background --- distance scale --- intergalactic
53: medium --- X-rays: galaxies: clusters}
54:
55:
56: \section{INTRODUCTION}
57: \label{sec:intro}
58:
59: The majority of helium in the universe, predominantly in the form of $^4$He,
60: was produced during the Big Bang. For the \wmap\ value of $\Omega_b
61: h^2=0.0223\pm 0.0007$ (Spergel et al.\ 2007), the standard hot big bang
62: nucleosynthesis model predicts a primordial fraction of helium in the total
63: baryonic mass density of $Y=0.2482\pm 0.0007$ (Walker et al.\ 1991; Kneller
64: \& Steigman 2004). Recent spectral measurements in metal-poor extragalactic
65: H{\small II} regions give a value within 1\% of this theoretical prediction,
66: with similarly small uncertainties (Izotov et al.\ 2007; Peimbert et al.\
67: 2007). Thus, the primordial helium abundance appears to be known quite
68: accurately.
69:
70: Helium abundance in the hot intracluster medium (ICM) may differ
71: significantly from the primordial one. First, additional helium comes from
72: the stars. The ratio of star mass to ICM mass is higher in the cluster
73: centers, so stellar enrichment will be stronger there. However, the mass of
74: the primordial helium in the ICM is comparable to the total stellar mass in
75: a cluster, so helium enrichment by stars should not be significant (unlike
76: stellar contribution of heavier elements, which are present in the ICM in
77: trace amounts). A much greater increase of helium abundance in the central
78: regions of clusters may be caused by sedimentation of heavy nuclei of the
79: ICM in the cluster gravitational potential (Fabian \& Pringle 1977; Rephaeli
80: 1978; Abramopoulos, Chanan, \& Ku 1981; Gilfanov \& Sunyaev 1984; Qin \& Wu
81: 2000; Chuzhoy \& Nusser 2003; Chuzhoy \& Loeb 2004; Ettori \& Fabian 2006).
82: The consensus of the recent works is that if sedimentation is not
83: suppressed, then in a hot cluster undisturbed for several gigayears, the
84: relative helium abundance can increase by a factor of 2 or more within
85: $r<0.2-0.3 r_{\rm 200}$, and even more at smaller radii (Chuzhoy \& Loeb
86: 2004). However, sedimentation can be inhibited by several mechanisms,
87: including tangled magnetic fields (which should also suppress diffusion and
88: thermal conduction in the ICM, as seems to be observed, e.g., Ettori \&
89: Fabian 2000; Vikhlinin et al.\ 2001; Markevitch et al.\ 2003), gas mixing by
90: cluster mergers and turbulence, and the formation of a cluster cool core
91: (Ettori \& Fabian 2006), because the diffusion rate is a strong function of
92: the temperature.
93:
94: While mergers and turbulence should inhibit any contemporary sedimentation,
95: they are unlikely to permanently erase a large-scale abundance gradient
96: already present by the time of the disturbance. The reason is the ICM in
97: relaxed clusters is stratified, with low-entropy gas at the bottom of the
98: gravitational well and higher-entropy gas in the outskirts. Such a stable
99: gas distribution should restore itself, and any radial abundance gradient
100: with it, shortly after a disturbance, provided that small-scale ICM mixing
101: during a merger is inefficient. Indeed, we do observe radially declining
102: {\em iron}\/ abundance profiles in all relaxed clusters with sufficiently
103: detailed X-ray data, with the decline traced from the cluster cores to at
104: least $r\sim 0.5 r_{200}$ (e.g., Fukazawa et al.\ 1994; Tamura et al.\ 2004;
105: Vikhlinin et al.\ 2005). While the iron abundance gradients are probably
106: caused by enrichment rather than sedimentation, they should be old enough to
107: have survived a merger or two, suggesting that it is difficult to erase an
108: abundance gradient. Thus, given the uncertainties in the processes that
109: inhibit sedimentation in the ICM, it is unclear how significant it would be
110: in a typical cluster.
111:
112: As pointed out by Qin \& Wu (2000) and in later works (and summarized in
113: \S\ref{sec:fgas} below), helium abundance affects cluster quantities derived
114: from X-ray observations, such as the cluster total mass, gas mass, and gas
115: mass fraction (see also Belmont et al.\ 2005 for an application to the hot
116: gas in the center of our Galaxy). The assumed helium abundance also affects
117: abundances of heavier elements derived from their X-ray emission lines
118: (e.g., Drake 1998; Ettori \& Fabian 2006). Most of the current X-ray
119: cluster analyses are restricted to bright central regions --- precisely
120: those regions that may be affected by sedimentation. Unfortunately, helium
121: in the ICM is fully ionized and not directly observable by spectroscopic
122: means. For this reason, its abundance is unknown and has to be adopted from
123: unrelated measurements, e.g., helioseismology (for a review see, e.g.,
124: Lodders 2003). The widely used X-ray spectral fitting package {\small
125: XSPEC} offers a choice of abundance models with the number density of
126: helium relative to hydrogen,
127: %
128: \begin{equation}
129: x\equiv \frac{n_{\rm He}}{n_p},
130: \end{equation}
131: %
132: spanning a range between 0.0792 (Lodders 2003) to 0.0977 (Anders \& Grevesse
133: 1989 and some others). For comparison, if one takes the abundances of
134: heavier elements to be $0.3-0.5$ solar (as in clusters), the CMB-based
135: primordial helium abundance (Spergel et al.\ 2007) corresponds to $x=0.083$.
136:
137: The unknown cluster helium abundance is a source of uncertainty for X-ray
138: cluster-based cosmology studies. In this Letter, I propose a way to measure
139: it.
140:
141:
142: \section{X-RAY DERIVED QUANTITIES}
143:
144: \subsection{Gas mass fraction}
145: \label{sec:fgas}
146:
147: As mentioned above, a helium abundance in the ICM is implicitly assumed in
148: most X-ray derived cluster quantities, such as gas mass, total mass from the
149: hydrostatic equilibrium equation, and their ratio $f_{\rm gas}$. All of
150: these are being used for cosmological tests (see, e.g., Vikhlinin et al.\
151: 2003, Henry 1997, and Allen et al.\ 2004 for the three quantities,
152: respectively), with projects underway to use them for ``precision
153: cosmology''. Below, their dependences on the assumed helium abundance are
154: written in a form relevant for the X-ray analysis (that is, fixing the X-ray
155: and other observables). For simplicity, I assume a uniform abundance over
156: the region involved. Current cosmological tests use relatively small
157: central cluster regions ($r<r_{500}$ or even $r<r_{2500}$), for which the
158: effects of sedimentation and enrichment can be significant.
159:
160: In a fully ionized intracluster plasma, the number of electrons per proton
161: is
162: %
163: \begin{equation}
164: \frac{n_e}{n_p} =1+2x+x_{eh} \approx 1+2x,
165: \end{equation}
166: %
167: where $x_{eh}$ represents electrons from elements heavier than helium;
168: $x_{eh}\approx 0.005$ for the intracluster chemical abundances ($0.3-0.5$
169: solar), and we will neglect it. The mean molecular weight of the ICM is
170: %
171: \begin{equation}
172: \mu = \frac{1+4x+m_h}{2+3x+x_{eh}+x_{h}} \approx \frac{1+4x}{2+3x},
173: \label{eq:mu}
174: \end{equation}
175: %
176: where $m_h$ and $x_h$ are the mass and number density contribution from
177: elements heavier than helium. For clusters, $m_h\approx 0.01$ and can be
178: neglected; $x_{h}\ll x_{eh}$ and is certainly negligible as well. Thus, for
179: clarity, we will ignore heavy elements below.
180:
181: The total mass of a cluster within a certain radius, derived under the
182: assumption of hydrostatic equilibrium (e.g., Sarazin 1988) and (for our
183: illustrative purposes) isothermality, is
184: %
185: \begin{equation}
186: M_{\rm tot} \propto \frac{T_e}{\mu} \frac{d \log \rho_{\rm gas}}{d \log r}
187: \propto \frac{1}{\mu} = \frac{2+3x}{1+4x}.
188: \label{eq:mtot}
189: \end{equation}
190: %
191: Here we used the fact that the logarithmic density gradient does not depend
192: on $x$\/ under our assumption that helium abundance is spatially uniform;
193: this relation will be more complicated if one uses a range of radii where
194: this assumption does not hold. The electron temperature $T_e$, derived from
195: the shape of the X-ray spectrum, is practically independent of the helium
196: abundance.
197:
198:
199: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
200: \begin{figure}[b]
201: \vspace*{5mm}
202: \center
203: \includegraphics[width=0.45\textwidth, bb=31 174 536 645]%
204: {f1.eps}
205:
206: \caption{The relative error in the X-ray derived cluster gas mass fraction,
207: $f_{\rm gas}$, as a function of the error in the cluster helium abundance,
208: $x$ (see eq.\ \ref{eq:fgas}).}
209:
210: \label{fig:fgas}
211: \end{figure}
212: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
213:
214: Let us now consider the cluster gas mass. Provided the cluster absolute size
215: is known, it is proportional to the plasma density. In X-ray analysis, the
216: plasma hydrogen density $n_p$ is derived from the normalization of the
217: cluster continuum spectrum (assuming that emission lines from heavy elements
218: are detected and properly modeled). The X-ray continuum luminosity is
219: %
220: \begin{equation}
221: L_X \propto n_e n_p \epsilon_{ep} (1+4x)
222: = n_p^2 \epsilon_{ep} (1+4x)(1+2x)
223: \end{equation}
224: %
225: where $\epsilon_{ep}$ is bremsstrahlung emissivity for a pure
226: electron-proton plasma, and the factor $(1+4x)$ accounts for the additional
227: bremsstrahlung on helium nuclei with charge 2. Again, heavier elements add
228: very little to the continuum emission. $L_X$ is the observable quantity that
229: one obtains directly from spectral fitting (and a known distance). Fixing
230: it, we obtain the dependence of the derived $n_p$ on the assumed helium
231: abundance:
232: %
233: \begin{equation}
234: n_p \propto [(1+4x)(1+2x)]^{-1/2}.
235: \label{eq:np}
236: \end{equation}
237: %
238: The gas mass then depends on $x$\/ as follows:
239: %
240: \begin{equation}
241: M_{\rm gas} \propto n_p + 4n_{\rm He} = n_p (1+4x)
242: \propto \left(\frac{1+4x}{1+2x}\right)^{1/2}.
243: \label{eq:mgas}
244: \end{equation}
245: %
246: Note that $M_{\rm gas}$ and $M_{\rm tot}$ change with $x$\/ in the opposite
247: directions, so their ratio, the gas mass fraction, depends on $x$\/ stronger
248: than either of these quantities:
249: %
250: \begin{equation}
251: f_{\rm gas}=\frac{M_{\rm gas}}{M_{\rm tot}}
252: \propto \frac{(1+4x)^{3/2}}{(2+3x)(1+2x)^{1/2}}.
253: \label{eq:fgas}
254: \end{equation}
255: %
256: We do not know the true value of $x$\/ and have to assume one to calculate
257: $f_{\rm gas}$; Fig.\ \ref{fig:fgas} shows the resulting relative error.
258: Note that the above $f_{\rm gas}(x)$ dependence is weaker than that derived
259: by Ettori \& Fabian (2006); the difference is due to our fixing of the
260: observable quantity $L_X$ when deriving $n_p$ in order to mimic the X-ray
261: data analysis.
262:
263: Fig.\ \ref{fig:fgas} shows that an error by a factor of 2 in the assumed
264: helium abundance corresponds to a $\sim 15$\% error in $f_{\rm gas}$. This
265: is comparable to the expected difference between the apparent $f_{\rm gas}$
266: at $z=1$ for open and flat $\Omega_m=0.3$ cosmologies. Allen et al.\ (2004)
267: derived $f_{\rm gas}$ values within $r<r_{2500} \approx 0.25 r_{200}$ for a
268: sample of hot clusters, detected such a difference, and used it as evidence
269: for $\Omega_\Lambda>0$. Of course, this cosmological test is based on a
270: comparison of $f_{\rm gas}$ at high and low redshifts, so the value of
271: helium abundance does not matter as long as it does not evolve between those
272: redshifts. Furthermore, the sign of the error arising from wrongly assuming
273: a primordial helium abundance in the presence of sedimentation is such that
274: the derived $f_{\rm gas}$ would be closer to its ``pre-sedimentation''
275: value, which is what one would ideally want to use for such a test (A.
276: Vikhlinin, private communication). A discussion of this error-cancellation
277: effect is beyond the scope of this paper (it would require a more accurate
278: calculation than that used for eq.\ \ref{eq:mtot}). However, it is clear
279: that any tests relying on an even higher accuracy of $f_{\rm gas}$ (such as
280: deriving the dark energy equation of state, which would need a few percent
281: accuracy on $f_{\rm gas}$) will require the knowledge of the intracluster
282: helium abundance at different redshifts. We will return to this in
283: \S\ref{sec:da}.
284:
285: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
286: \begin{figure}
287: \center
288: \includegraphics[width=0.45\textwidth, bb=31 174 536 645]%
289: {f2.eps}
290:
291: \caption{The apparent $H_0$ value from a combination of SZ
292: and X-ray cluster data, as a function of the error in the cluster helium
293: abundance, $x$ (see eq.\ \ref{eq:da}), shown by solid line. For
294: illustration, we take {\em WMAP}'s $H_0=73$ \kmsmpc\ as a ``true'' value
295: (horizontal black dashed line; Spergel et al.\ 2007). White dashed line
296: and gray error band (statistical only, 68\%) overlays the $H_0$
297: measurement by Bonamente et al.\ (2006), who assumed a helium abundance
298: from Anders \& Grevesse (1989).}
299:
300: \label{fig:h0}
301: \vspace{5mm}
302: \end{figure}
303: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
304:
305:
306: \subsection{Hubble constant}
307: \label{sec:h0}
308:
309: It has long been suggested (Cavaliere, Danese, \& de Zotti 1979; Silk \&
310: White 1978) that combining the X-ray and Sunyaev-Zeldovich (SZ) observations
311: of a cluster can be used to measure the absolute distance to the cluster.
312: This method uses the fact that the SZ decrement and the X-ray brightness
313: depend on different powers of the intracluster electron density, whose value
314: can thus be determined and converted to the distance. Below I show how this
315: measurement depends on the assumed helium abundance. The underlying reason
316: for this dependence is that the SZ effect is caused only by electrons, while
317: the X-ray emission is caused by scattering of electrons on protons and
318: helium nuclei. The fact that helium sedimentation would affect the cluster
319: SZ decrement was mentioned by Gilfanov \& Sunyaev (1984) but, to our
320: knowledge, never considered in any SZ data analyses.
321:
322: In a nonrelativistic approximation, the SZ signal in the direction of a
323: cluster is proportional to the comptonization parameter
324: %
325: \begin{equation}
326: y \equiv \int \frac{kT_e}{m_e c^2}\,\sigma_T\, n_e(l)\, dl \,\propto\, T_e\,
327: n_{e0}\, d_a
328: \end{equation}
329: %
330: (Sunyaev \& Zeldovich 1972), where the integral is along the line of sight,
331: $n_{e0}$ is some characteristic electron density (e.g., near the cluster
332: center), and $d_a$ is the angular distance to the cluster. Here for the
333: rightmost part of the equation, the plasma cloud is assumed isothermal and
334: spherically symmetric. The latter assumption is important (the
335: line-of-sight distribution of the gas density is taken to be the same as
336: that in the plane of the sky) but cannot be tested directly, so in practice,
337: a cluster sample has to be used to average out any possible ellipticities.
338: The surface brightness of the X-ray continuum emission from the same line of
339: sight is given by
340: %
341: \begin{equation}
342: S_X = \int n_e(l)\, n_p(l)\, \epsilon_{ep}\, (1+4x)\, dl \,\propto\,
343: n_{e0}^2\, d_a\, \frac{1+4x}{1+2x}.
344: \end{equation}
345: %
346: The distance can be determined by combining the above equations as follows:
347: %
348: \begin{equation}
349: d_a \propto \frac{y^2}{S_X\,T_e^2}\, \frac{1+4x}{1+2x}.
350: \label{eq:da}
351: \end{equation}
352: %
353: The quantities $y$, $S_X$ and $T_e$ are directly measured, but helium
354: abundance $x$\/ has to be assumed. Fig.\ \ref{fig:h0} shows the effect of
355: this assumption on the derived $d_a$ (or $H_0\propto d_a^{-1}$), based on
356: eq.\ (\ref{eq:da}). For example, a factor of 2 error in helium abundance
357: results in a 10\% distance error.
358:
359:
360: \section{Measuring helium abundance in ICM}
361:
362: The above dependence of the SZ--X-ray distances on the helium abundance can
363: be turned around and used to {\em derive}\/ the cluster helium abundance ---
364: provided the distance scale is known independently, for example, from
365: Cepheids and supernovae (Freedman et al.\ 2001; Sandage et al.\ 2006) or
366: from the CMB fluctuations (Spergel et al.\ 2007). At present, the latter
367: two methods yield similar or smaller uncertainties on distances than those
368: from state of the art SZ--X-ray studies (e.g., Bonamente et al.\ 2006).
369: This is because measurement errors on the SZ signal are still quite large,
370: while this observable ($y$) enters squared in eq.\ (\ref{eq:da}). Neither
371: the supernovae distances nor those from CMB fluctuations have any
372: significant dependence on primordial helium abundance (Ichikawa \& Takahashi
373: 2006).
374:
375: Fig.\ \ref{fig:h0} shows the value of $H_0=77.6^{+4.8}_{-4.3}$ \kmsmpc\ from
376: Bonamente et al.\ (2006), derived from the SZ and X-ray data on 38 clusters
377: at different redshifts. Of their 3 reported values, I chose the least
378: model-dependent one, without the hydrostatic equilibrium assumption and
379: excluding the central cool regions. The above error bars are 68\%
380: statistical-only, to illustrate an accuracy not quite achieved yet but
381: within immediate reach (once the instruments have been better calibrated);
382: their current systematic uncertainties are twice as big.
383:
384: According to M. Bonamente (private communication), they assumed the Anders
385: \& Grevesse (1989) helium abundance, which is a factor of 1.17 higher than
386: the primordial value (\S\ref{sec:intro}). Had they used the primordial
387: abundance, their result would be $H_0\simeq 79$ \kmsmpc. If, for the sake
388: of argument, we take the \wmap\ value of $H_0$ as ``true'', and attribute
389: the difference between these values to a helium abundance error (Fig.\
390: \ref{fig:h0}), we conclude that it should be a factor $1.9\pm0.8$ higher
391: than the primordial value. Of course, at this accuracy, it is consistent
392: with the primordial value, but we can already exclude some of the more
393: extreme predictions for helium sedimentation.
394:
395: This and many other current SZ experiments use interferometric mapping of
396: the radio brightness. For typical clusters, this experimental design
397: effectively ``subtracts'' the signal from the shell outside $r\sim 1$ Mpc,
398: so the above constraint corresponds to this approximate central region.
399: This is about the same radius as used for the $f_{\rm gas}$ test by Allen et
400: al.\ (2004).
401:
402:
403: \section{SUMMARY AND DISCUSSION}
404: \label{sec:disc}
405:
406: The usual assumption of a primordial helium abundance for the ICM is not
407: necessarily correct, especially in the cluster central regions where helium
408: may concentrate via several mechanisms, such as sedimentation. We cannot
409: measure the cluster helium abundance spectroscopically. As pointed out by
410: many authors, an incorrect helium abundance assumption will result in
411: incorrect cluster gas masses, total masses and gas mass fractions derived
412: from the X-ray data.
413:
414: In this paper, we point out that cluster distances derived using the
415: SZ--X-ray combination also depend on helium abundance. If one gives up on
416: the original purpose of this method and combines it with an independent
417: distance scale estimator (such as supernovae or CMB fluctuations), one can
418: take advantage of this dependence and derive the cluster helium abundance.
419: At present, this seems to be the only practical way of measuring it. The
420: accuracy may already be interesting --- if one compares the best SZ--X-ray
421: value for $H_0$ (Bonamente et al.\ 2006) with the CMB value (Spergel et al.\
422: 2007), the difference corresponds to a helium abundance $1.9\pm0.8$ times
423: the primordial value within the cluster central 1 Mpc regions (68\%
424: statistical-only uncertainty). Increasing the number of clusters in the
425: sample and improving the SZ data accuracy will reduce this uncertainty.
426:
427: Chuzhoy \& Loeb (2004) proposed another way of constraining the cluster
428: helium abundance --- by comparing the cluster total masses derived using the
429: hydrostatic equilibrium assumption (eq.\ \ref{eq:mtot} above) and any other
430: technique independent of helium abundance, such as gravitational lensing.
431: However, this approach appears less practical at present, because the
432: expected apparent mass difference ($\sim 10$\% for a factor of 2 error in
433: helium abundance) is much smaller than persistent discrepancies between
434: these mass estimators that are likely to be caused by substructure in the
435: dark matter distribution and deviations from hydrostatic equilibrium in the
436: ICM (e.g., Gavazzi 2005; Meneghetti et al.\ 2007). In comparison, the
437: SZ--X-ray method uses the same object --- the ICM --- at both wavelengths
438: and does not rely on hydrostatic equilibrium. It does, however, assume that
439: the ICM is not clumpy (which appears to be supported by \chandra\ imaging),
440: and requires accurate mapping of the ICM temperature structure. It also
441: needs spherical symmetry, which has always been an issue for the SZ method
442: of $H_0$ determination. It can be overcome by proper (i.e., X-ray) selection
443: of a sample of relaxed clusters that is also big enough to average out the
444: asymmetries.
445:
446: Chuzhoy \& Loeb (2004) also pointed out that a higher helium abundance in
447: the ICM would affect stars that forms out of this ICM. Thus, evidence of
448: helium sedimentation may also be found in the spectra of the central cluster
449: galaxies.
450:
451:
452: \subsection{The $d_a(z)$ cosmological test and helium abundance}
453: \label{sec:da}
454:
455: The main reason why we want to know the cluster helium abundances is to
456: remove the related uncertainty from the cluster-based cosmological tests,
457: such as the growth of structure tests that use cluster mass functions and
458: the $d_a(z)$ tests that use $f_{\rm gas}$ or the SZ--X-ray distances.
459: Obviously, using a competing distance estimator to measure helium abundances
460: will introduce degeneracies into the resulting cosmological constraints, the
461: detailed analysis of which is beyond the scope of this paper. In principle,
462: the above two distance tests can be combined to solve for the dependence on
463: helium abundance, because, as seen from eqs.\ (\ref{eq:fgas}) and
464: (\ref{eq:da}), the distances derived from $f_{\rm gas}$ ($d_a\propto f_{\rm
465: gas}^{2/3}$) and from the SZ--X-ray method depend on $x$ differently. One
466: complication is that if one allows for helium sedimentation, the basic
467: assumption of the $f_{\rm gas}$-based test, that $f_{\rm gas}$ within a
468: certain central region does not evolve with $z$, may be violated. So in
469: practice, one would have to fit together the X-ray and SZ data for a sample
470: of relaxed clusters spanning a range of $z$, parameterize and fit any
471: systematic change of helium abundance with redshift (hopefully small or
472: negligible), and use modeling to deduce ``pre-sedimentation'' $f_{\rm gas}$
473: values (to within a scaling factor), which would be the ones proportional to
474: the universal baryon fraction. To derive cosmological parameters other than
475: $H_0$, the absolute values of $d_a$ (and so the absolute values of $x$) are
476: not needed, only its change with $z$. However, independent distances for at
477: least a few clusters will be required to determine if any helium
478: sedimentation occurs at all, and if so, to model it.
479:
480:
481: \acknowledgements
482:
483: I thank Alexey Vikhlinin, Daisuke Nagai, Avi Loeb, and Jeremy Drake for
484: useful discussions. Support for this work was provided by NASA contract
485: NAS8-39073 and \chandra\ grant GO6-7126X.
486:
487:
488: \begin{references}
489:
490: \reference{} Abramopoulos, F., Chanan, G.~A., \& Ku, W.~H.-M.\ 1981, \apj,
491: 248, 429
492:
493: \reference{} Anders, E., \& Grevesse, N. 1989, \gca, 53, 197
494:
495: \reference{} Belmont, R., Tagger, M., Muno, M., Morris, M., \& Cowley, S.\
496: 2005, \apjl, 631, L53
497:
498: \reference{} Cavaliere, A., Danese, L., \& de Zotti, G.\ 1979, \aap, 75,
499: 322
500:
501: \reference{} Chuzhoy, L., \& Loeb, A.\ 2004, \mnras, 349, L13
502:
503: \reference{} Chuzhoy, L., \& Nusser, A.\ 2003, \mnras, 342, L5
504:
505: \reference{} Drake, J.~J.\ 1998, \apjl, 496, L33
506:
507: \reference{} Ettori, S., \& Fabian, A.~C.\ 2000, \mnras, 317, L57
508:
509: \reference{} Ettori, S., \& Fabian, A.~C.\ 2006, \mnras, 369, L42
510:
511: \reference{} Fabian, A.~C., \& Pringle, J.~E.\ 1977, \mnras, 181, 5P
512:
513: \reference{} Freedman, W.~L., et al.\ 2001, \apj, 553, 47
514:
515: \reference{} Fukazawa, Y., Ohashi, T., Fabian, A.~C., Canizares, C.~R.,
516: Ikebe, Y., Makishima, K., Mushotzky, R.~F., \& Yamashita, K.\ 1994, \pasj,
517: 46, L55
518:
519: \reference{} Gavazzi, R. 2005, \aap, 443, 793
520:
521: \reference{} Gilfanov, M.~R., \& Sunyaev, R.~A.\ 1984, Soviet Astronomy
522: Letters, 10, 137
523:
524: \reference{} Ichikawa, K., \& Takahashi, T.\ 2006, \prd, 73, 063528
525:
526: \reference{} Izotov, Y.~I., Thuan, T.~X., \& Stasinska, G.\ 2007, ApJ, in
527: press; arXiv:astro-ph/0702072
528:
529: % SBBN abundance fits
530: \reference{} Kneller, J.~P., \& Steigman, G.\ 2004, New Journal of Physics,
531: 6, 117
532:
533: \reference{} Lodders, K. 2003, ApJ, 591, 1220
534:
535: \reference{} Markevitch, M., et al.\ 2003, \apjl, 586, L19
536:
537: \reference{} Meneghetti, M., et al. 2007, \aap, 461, 25
538:
539: \reference{} Peimbert, M., Luridiana, V., \& Peimbert, A.\ 2007, ApJ,
540: submitted; arXiv:astro-ph/0701580
541:
542: \reference{} Qin, B., \& Wu, X.-P.\ 2000, \apjl, 529, L1
543:
544: \reference{} Sandage, A., Tammann, G.~A., Saha, A., Reindl, B., Macchetto,
545: F.~D., \& Panagia, N.\ 2006, \apj, 653, 843
546:
547: \reference{} Sarazin, C. L. 1988, X-ray emission from clusters
548: of galaxies (Cambridge University Press)
549:
550: \reference{} Silk, J., \& White, S.~D.~M.\ 1978, \apjl, 226, L103
551:
552: \reference{} Spergel, D.~N., et al.\ 2007, ApJ in press, arXiv:astro-ph/0603449
553:
554: \reference{} Sunyaev, R. A., \& Zeldovich, Y. B. 1972, Comments on
555: Astrophys.\ and Space Phys., 4, 173
556:
557: \reference{} Tamura, T., Kaastra, J.~S., den Herder, J.~W.~A., Bleeker,
558: J.~A.~M., \& Peterson, J.~R.\ 2004, \aap, 420, 135
559:
560: \reference{} Vikhlinin, A., Markevitch, M., \& Murray, S.~S.\ 2001, \apj,
561: 551, 160
562:
563: \reference{} Vikhlinin, A., Markevitch, M., Murray, S.~S., Jones, C.,
564: Forman, W., \& VanSpeybroeck, L.\ 2005, \apj, 628, 655
565:
566: \reference{} Walker, T.~P., Steigman, G., Kang, H.-S., Schramm, D.~M., \&
567: Olive, K.~A.\ 1991, \apj, 376, 51
568:
569: \end{references}
570:
571: \end{document}
572: