1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 4 distribution.
4: % Version 4.0 of REVTeX, August 2001
5: %
6: % Copyright (c) 2001 The American Physical Society.
7: %
8: % See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: % 1) latex apssamp.tex
17: % 2) bibtex apssamp
18: % 3) latex apssamp.tex
19: % 4) latex apssamp.tex
20: %
21: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
22: \documentclass[onecolumn,useAMS,usenatbib]{mn2e}
23: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
24:
25: % Some other (several out of many) possibilities
26: %\documentclass[preprint,aps]{revtex4}
27: %\documentclass[preprint,aps,draft]{revtex4}
28: %\documentclass[prb]{revtex4}% Physical Review B
29:
30: \usepackage{graphics,epsfig}% Include figure files
31: \usepackage{times}% Include figure files
32: \usepackage{dcolumn}% Align table columns on decimal point
33: %\usepackage{bm}% bold math
34: %\usepackage{natbib}
35:
36: %\nofiles
37:
38: \newcommand{\al}{\ensuremath{\alpha}}
39: \newcommand{\dal}{\ensuremath{\Delta \alpha/ \alpha}}
40: \newcommand{\gp}{\ensuremath{{g}_{p}}}
41: \newcommand{\lqcd}{\ensuremath{\Lambda_{\rm QCD}}}
42: \newcommand{\y}{\ensuremath{{m}_{e}/{m}_{p}}}
43: \newcommand{\dy}{\ensuremath{\Delta {m_e}/{m_p}/{m_e}/{m_e}}}
44:
45:
46: \newcommand{\beq}{\begin{equation}}
47: \newcommand{\eeq}{\end{equation}}
48: \newcommand{\noi}{\noindent}
49: \newcommand{\lb}{\left(}
50: \newcommand{\rb}{\right)}
51: \newcommand{\lsb}{\left[}
52: \newcommand{\rsb}{\right]}
53:
54:
55: %\preprint{APS/123-QED}
56:
57: \title[The use of OH ``main'' lines to $\ldots$ ]{The use of OH ``main'' lines to constrain the variation of fundamental constants}
58: \author[Kanekar \& Chengalur]{Nissim Kanekar$^1$\thanks{E-mail: nissim@astro.rug.nl (NK);
59: chengalu@ncra.tifr.res.in (JNC);}, J. N. Chengalur$^2$\footnotemark[1] \\
60: $^{1}$ Kapteyn Institute, University of Groningen, Post Bag 800, 9700 AV Groningen, The Netherlands\\
61: $^{2}$ National Centre for Radio Astrophysics, Post Bag 3, Ganeshkhind, Pune 411 007, India }
62:
63: \begin{document}
64:
65: \date{Received mmddyy/ accepted mmddyy}
66:
67:
68: \maketitle
69:
70: \label{firstpage}
71:
72:
73:
74: \begin{abstract}
75: We describe a new technique to measure variations in the fundamental
76: parameters $\alpha$ and $y \equiv m_e/m_p$, using the sum of the frequencies
77: of cm-wave OH ``main'' lines. The technique is $\sim$ three orders of
78: magnitude more sensitive than that of \cite{chengalur03}, which
79: utilised only the four 18cm OH lines. The increase in sensitivity stems from the use
80: of OH ``main'' lines
81: arising from different rotational states, instead of the frequency difference
82: between lines from the same state. We also show that redshifts of the main
83: OH 18cm and 6cm lines can be combined with the redshift of an HCO$^+$ transition
84: to measure any evolution in $\alpha$ and $y$. Both 18cm main lines and a number
85: of HCO$^+$ lines have already been detected in absorption in four cosmologically distant
86: systems; the detection of the main 6cm OH line in any of these systems would thus
87: be sufficient to simultaneously constrain changes in $\alpha$ and $y$ between
88: the absorption redshift and today.
89:
90: \end{abstract}
91:
92: \begin{keywords}
93: Line: profiles -- techniques: spectroscopic -- radio lines
94: \end{keywords}
95: \maketitle
96:
97: \section{\label{sec:intro} Introduction}
98:
99: In recent times, quasar absorption lines have emerged as an excellent probe of
100: changes in the values of the fundamental ``constants''
101: (e.g.~\citealt{webb99,carilli00,ivanchik03}). Such variations are expected in theories
102: like extra-dimensional Kaluza-Klein models or super-string theories, where
103: values of the coupling parameters such as the fine structure constant $\alpha$ or
104: the gravitational constant $G$ depend on the expectation values of some cosmological
105: scalar field(s); changes in these parameters are thus to be expected if the latter
106: varies with space and/or time. Various experimental and observational bounds are
107: available on the temporal evolution of different coupling constants: these include
108: the fine structure constant, $\al$ \citep{ivanchik99,webb01}, the
109: gravitational constant $G$ \citep{teller48,damour91}, the combination $g_p \al^2$
110: (where $g_p$ is the proton g-factor; \citealt{drinkwater98,carilli00}), the ratio
111: of electron mass to proton mass $y \equiv m_e/m_p$ \citep{ivanchik03}, etc.
112: \citet{uzan03} provides a review of the available measurements.
113:
114: The most interesting of the new astrophysical estimates are the recent work
115: of \cite{webb99,webb01} who claim a detection of changes in the
116: numerical value of the fine structure constant $\al$ between high redshift,
117: $z \sim 3.5$, and the present epoch. The authors initially applied a new
118: `many-multiplet' method to absorbers with $1 \la z \la 1.6$ to estimate
119: $\dal = (-1.88 \pm 0.53) \times 10^{-5}$ between redshifts $z \sim 1.6$ and
120: today \citep{webb99}. This was followed by the use of this method to
121: estimate $\dal = (-0.72 \pm 0.18) \times 10^{-5}$ over the redshift
122: range $0.5 < z < 3.5$ \citep{webb01} (see, however, \citealt{bekenstein03})
123: On the other hand, \cite{ivanchik03} constrain the variation in $m_e/m_p$
124: to be $(3.0 \pm 2.4) \times 10^{-5}$ over a similar redshift range ($ 0< z <3$),
125: comparable to the change claimed in the fine structure constant (albeit using a
126: different absorber sample). This is somewhat
127: surprising, given that most of the above theoretical analyses expect changes
128: in different fundamental constants to be coupled: for example, \cite{calmet02}
129: and \cite{langacker02} find that variations in the value of $\al$ should be
130: accompanied by much larger changes
131: (by $\sim 2$ orders of magnitude) in the value of $m_e/m_p$.
132:
133: We have recently (\citealt{chengalur03}; hereafter Paper I) demonstrated a new technique to measure
134: (or constrain) changes in the fundamental constants using 18cm OH lines. This
135: method uses the fact that the four OH lines arise from two very different physical
136: phenomena, $\Lambda$-doubling and hyperfine structure, and thus have different
137: dependences on the parameters $\alpha$, $y$ and $g_p$. Observations
138: of all four OH 18cm transitions in a single cosmologically distant absorber can
139: thus be used to simultaneously estimate variations in $y$ and $\alpha$,
140: assuming that the proton g-factor remains unchanged (e.g.~\citealt{webb01,carilli00}).
141: We have also used the linear relationship between OH and HCO$^+$ column densities
142: observed both in the Milky Way and in four molecular absorbers out to $z \sim 1$
143: to argue that HCO$^+$ and OH lines probably arise from the same spatial location
144: and are thus unlikely to have velocity offsets relative to each other. The OH 18cm
145: redshifts can then be combined with the redshift of a single HCO$^+$ line to
146: simultaneously estimate changes in all three fundamental parameters $\alpha$, $g_p$
147: and $y$, {\it in the same object}.
148:
149: A problem with the above approach is that two of the equations used in the
150: analysis (equations~(10) and (11) in Paper~I) involve the separation
151: of two line frequencies. The four 18cm lines have rest frequencies of 1665.4018~MHz
152: and 1667.3590~MHz (``main'' lines, with $\Delta F = 0$), and 1612.2310~MHz and
153: 1720.5299~MHz (``satellite'' lines, with $\Delta F = 1$). The separation between the
154: main line frequencies is a factor of $\sim 1600$ smaller than the sum of these frequencies
155: while the separation between the satellite frequencies is around 30 times smaller
156: than the above sum. This implies that the error on the redshift of the two
157: frequency differences is worse than the error on the redshift of the sum of
158: main line frequencies by the same factors and this large error propagates
159: into the estimates of changes in the values of $\alpha$, $y$ and $g_p$. To
160: emphasise this point, we note that the error on the sum of the main
161: line redshifts in B0218+357 is $\sim 5.6 \times 10^{-6}$, while that on the difference
162: between these redshifts is $\sim 6.7 \times 10^{-3}$ (Paper I); it is the latter error
163: that dominates the accuracy of the technique in constraining any changes in
164: the fundamental parameters (and which hence resulted in the large errors
165: in the analysis of the OH, HCO$^+$ and HI lines from the $z \sim 0.6846$ absorber
166: towards B0218+357 in Paper~I).
167:
168: We describe in this Letter a new approach to simultaneously measure changes
169: in $\alpha$ and $y$ using OH ``main'' lines arising from different OH rotation
170: states. This is based on the fact that the exact $\Lambda$-doubling
171: frequency split depends on the specific quantum numbers of the state in question
172: and has a different dependence on $\alpha$ and $y$ in each state. Since the method
173: only uses the sum of different ``main'' line frequencies, it is far
174: more sensitive than that discussed above (Paper~I), which also uses the difference
175: between pairs of measured frequencies. Further,
176: the sum of OH main line frequencies does not depend on the proton g-factor
177: (as the hyperfine effects cancel out); main lines from any three OH rotation
178: states can thus be used to simultaneously measure changes in $y$ and $\alpha$.
179: We also show that the main lines from two OH rotation states can be used in
180: conjunction with an HCO$^+$ transition to simultaneously measure variations in
181: $\alpha$ and $y$. The use of OH main lines to measure changes in the fundamental
182: constants has also been discussed by \cite{darling03}; this analysis was, however,
183: based on a simpler approximation to the OH energy levels and also only considered
184: variations in the fine structure constant $\alpha$.
185:
186: \section{The sum of OH ``main'' line frequencies}
187:
188: The sum of the ``main'' line frequencies $\nu_s$ in an OH rotation state
189: essentially gives the energy split due to $\Lambda$-doubling; this can be
190: written as
191:
192: \beq
193: \label{eqn:sum1}
194: \nu_s\;\; = \;\; q_\Lambda\lb J+1/2 \rb\lsb
195: \lb 2 + \frac{A'}{B'} \rb \lb 1 + \frac{2 - A/B}{X} \rb +
196: \frac{4 \lb J + 3/2 \rb \lb J - 1/2 \rb}{X} \rsb \;\; ,
197: \eeq
198:
199: \noi where $J$ is the rotational quantum number, $A$, the fine structure interaction
200: constant and $B$, the rotation constant \citep{vanvleck29,townes55}. The quantity $X$ is
201: defined by
202: \beq
203: \label{eqn:x}
204: X = \pm \lsb \lb A/B \rb \{ {\lb A/B\rb} - 4 \}+ 4\lb J + 1/2 \rb^2 \rsb^{1/2} \;\; ,
205: \eeq
206: \noi with the negative sign for the $^2\Pi_{3/2}$ state and the positive sign for
207: the $^2\Pi_{1/2}$ state \citep{townes55}. Further, $q_\Lambda \approx 4B^2/h\nu_e$,
208: where $h\nu_e$ is the energy difference between the ground and first excited
209: electronic state. Numerically, $A/B = -7.547$ and $A'/B' = -6.073$ \citep{townes55}.
210: The above quantities have the following dependences on the fundamental constants
211: $\alpha$, $y$ and $R_\infty$ : $A' \propto A \propto \alpha^2 R_\infty$,
212: $B' \propto B \propto y R_\infty$, where $R_\infty$ is the Rydberg constant.
213: For the rotation constant $B$,
214: we have assumed, as usual (e.g.~\citealt{murphy01}), that variations in $(m_p/M)$,
215: which are suppressed by a factor $m_p/U \sim 100$ (where $M$ is the reduced mass and
216: U the binding energy) can be ignored. We thus have $ \lsb A'/B' \rsb \propto \lsb A/B
217: \rsb \propto \lb \alpha^2/y \rb $. Finally, we note that equation~(\ref{eqn:sum1})
218: does not depend on the proton g-factor $g_p$.
219:
220: Replacing the above scalings in equation~(\ref{eqn:sum1}) for $\nu_s$, we obtain
221: $\nu_s \propto y^2 R_\infty F\lb \alpha^2/y\rb \;\;$, where $F \equiv F \lb \beta \rb $
222: is a function which depends only on the ratio $\beta \equiv A/B \propto \alpha^2/y$ and
223: is defined by
224: \beq
225: F\lb \beta \rb = \lsb \lb 2 + \frac{6.073}{7.547}\beta \rb \lb 1 +
226: \frac{2 - \beta}{X\lb\beta\rb} \rb + \frac{4 \lb J + 3/2 \rb \lb J - 1/2 \rb}{X\lb \beta\rb} \rsb \; \; .
227: \eeq
228:
229: \noi Thus, the ratio of the change in the sum of any two main line frequencies
230: $\Delta \nu_s$ to their sum today is given by
231: \begin{eqnarray}
232: \frac{\Delta \nu_s }{\nu_s} &=& 2\frac{\Delta y}{y} + \frac{\Delta R_\infty}{R_\infty}
233: + \frac{\Delta F \lb \beta \rb} {F\lb\beta \rb}\\
234: \label{eqn:sum}
235: &=& 2\frac{\Delta y}{y} + \frac{\Delta R_\infty}{R_\infty}
236: + \frac{\beta}{F} \frac{dF}{d\beta} \lsb 2 \frac{\Delta \alpha}{\alpha} -
237: \frac{\Delta y}{y} \rsb
238: \end{eqnarray}
239:
240: \beq
241: \mathrm{where} \;\;\;\; \frac{dF}{d\beta} = C \lb 1 + \frac{2 - \beta}{X} \rb
242: - \frac{2 + C\beta}{X}\lsb 1 + \frac{2 - \beta}{X} \frac{dX}{d\beta} \rsb
243: - \frac{4 \lb J + 3/2 \rb \lb J - 1/2 \rb}{X^2} \frac{dX}{d\beta} \;\;.
244: \eeq
245: \noi In the above, $dX/d\beta = \lb \beta - 2\rb/X$ and $C = (6.073/7.547)$.
246: \noi Equation~(\ref{eqn:sum}) has been evaluated for the main lines of the
247: $^2\Pi_{3/2}, J = 3/2$ state in Paper~I; we
248: extend this analysis to the main lines of other rotational states that have been
249: detected in the Galactic interstellar medium. The results are listed below,
250: along with the rest frequencies of the ``main'' lines for each state :
251:
252: \begin{enumerate}
253: \item{$\mathbf{^2\Pi_{3/2}, J = 3/2}$ : Rest frequencies : 1667.3590 and 1665.4018~MHz
254: (i.e. $\sim $ 18cm). Equation~(\ref{eqn:sum}) yields
255: \beq
256: \label{eqn:18cm}
257: \frac{\Delta \nu_s }{\nu_s} = 2.571 \frac{\Delta y}{y} -
258: 1.141 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
259: \eeq
260: }
261: \item{$\mathbf{^2\Pi_{1/2}, J = 1/2}$ : This state has a single main line, at a rest
262: frequency of 4750.656~MHz ($\sim $ 6cm); the second main line corresponds to
263: an F = 0--0 transition, which is forbidden by the selection rules. This might appear
264: to be a drawback, since it would seem impossible to detect both main lines. However,
265: it turns out that the frequency separation between the two main lines vanishes, to first
266: order. This can be seen as the separation between the two main line frequencies
267: $\nu_{1}$ and $\nu_{2}$ is
268: \beq
269: \label{eqn:diff1}
270: \Delta \nu\;\; \equiv\;\; \nu_{1} - \nu_{2}\;\; =\;\; \frac{-2d \lb -X - 2 + \beta \rb }{3 X}
271: \eeq
272: \noi where $d$ is a hyperfine constant (equation~(46) of \citealt{dousmanis55}). Further,
273: $X = \beta -2$ for $J = 1/2$, implying that $\Delta \nu = 0$. Thus, the sum of the two
274: main line frequencies for the $J = 1/2$ case is, to first order, merely equal to twice
275: the F = 1--1 frequency and a detection of this transition is sufficient to measure
276: the above sum. In this case, equation~(\ref{eqn:sum}) yields
277: \beq
278: \label{eqn:6cm}
279: \frac{\Delta \nu_s }{\nu_s} = 0.509 \frac{\Delta y}{y} +
280: 2.982 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
281: \eeq
282: }
283: \item{$\mathbf{^2\Pi_{3/2}, J = 5/2}$ : Rest frequencies : 6030.747~MHz and 6035.092~MHz
284: ($\sim $ 5cm). Equation~(\ref{eqn:sum}) yields
285: \beq
286: \label{eqn:5cm}
287: \frac{\Delta \nu_s }{\nu_s} = 2.452 \frac{\Delta y}{y} -
288: 0.903 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
289: \eeq
290: }
291: \item{$\mathbf{^2\Pi_{1/2}, J = 3/2}$ : Rest frequencies : 7761.747~MHz and 7820.125~MHz
292: ($\sim $ 3.8cm). Equation~(\ref{eqn:sum}) yields
293: \beq
294: \label{eqn:3.8cm}
295: \frac{\Delta \nu_s }{\nu_s} = 0.072 \frac{\Delta y}{y} +
296: 3.857 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
297: \eeq
298: }
299: \item{$\mathbf{^2\Pi_{1/2}, J = 5/2}$ : Rest frequencies : 8135.870~MHz and 8159.587~MHz
300: ($\sim $ 3.7cm). Equation~(\ref{eqn:sum}) yields
301: \beq
302: \label{eqn:3.7cm}
303: \frac{\Delta \nu_s }{\nu_s} = -0.920 \frac{\Delta y}{y} +
304: 5.840 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
305: \eeq
306: }
307: \item{$\mathbf{^2\Pi_{3/2}, J = 7/2}$ : Rest frequencies : 13434.596~MHz and
308: 13441.4173~MHz ($\sim $ 2.2cm). Equation~(\ref{eqn:sum}) yields
309: \beq
310: \label{eqn:2.2cm}
311: \frac{\Delta \nu_s }{\nu_s} = 2.334\frac{\Delta y}{y} -
312: 0.678 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
313: \eeq
314: }
315: \end{enumerate}
316:
317: We note that each of the above equations~(\ref{eqn:18cm} -- \ref{eqn:2.2cm}) has
318: the same dependence on the Rydberg constant $R_\infty$, but different dependences on
319: $y$ and $\alpha$. If we have two transitions whose rest frequencies $\nu_1(0)$ and
320: $\nu_2(0)$ depend on redshift, due to the evolution of various fundamental
321: constants such as $\al$, $y$, etc, the first order difference between the measured
322: redshifts is (e.g.~Paper~I)
323: \beq
324: \label{eqn:redshift}
325: \frac {\Delta z }{ 1 + {\bar z} }\;\; =\;\; \lsb \frac{ \Delta \nu_2}{\nu_2(0)} \rsb -
326: \lsb \frac{\Delta \nu_1}{\nu_1(0)} \rsb \;\; ,
327: \eeq
328:
329: \noi where $\bar z$ is the mean measured redshift. Given two spectral lines (or
330: linear combinations of line frequencies) with different dependences on some
331: fundamental parameter, the differences between the measured redshifts can be used
332: to constrain the evolution of the parameter in question. Clearly, any three
333: of the equations~(\ref{eqn:18cm} -- \ref{eqn:2.2cm}) can be combined in pairs
334: in equation~(\ref{eqn:redshift}) to yield two simultaneous equations
335: in $\Delta \alpha / \alpha$ and $\Delta y / y$, which can then be solved to
336: measure any changes in both these quantities. Thus, the detection of the ``main''
337: OH lines in any three of the OH rotation states in a single absorber can be used
338: to simultaneously measure changes in $\alpha$ and $y \equiv m_e/m_p$ at the same
339: physical space-time location. Note that the large number of OH rotational
340: states allows a simple self-consistency check of any such measurement, by a search
341: for the ``main'' lines from a fourth OH rotation state. Since this technique
342: uses different lines from the same species, systematic velocity offsets between
343: the different absorption lines are unlikely to be the dominant source of error.
344: It should be noted, however, that the higher OH rotational levels may be
345: excited in regions with very different physical conditions from regions
346: giving rise to OH ground state absorption; the possibility of velocity
347: offsets between the 18cm absorbing gas and the 6cm or 5cm absorbing gas
348: hence cannot be ruled out. However, the OH column density of the absorbing
349: gas can be independently estimated from the different OH lines; if these estimates
350: are found to be in agreement, it would argue that all the lines originated in the same gas cloud.
351:
352: HCO$^+$ rotational transitions also have no dependence on the proton
353: g-factor $g_p$, with line frequencies proportional to $y R_\infty$. Moreover,
354: as discussed in Paper~I, the linear relation between HCO$^+$ and OH
355: column densities observed both in the galaxy \citep{liszt96} and out to $z \sim 1$
356: \citep{kanekar02} (extending over more than two orders of magnitude in column
357: density) suggests that the two species are likely to be located in the same
358: region of a molecular cloud. This implies that one of the OH rotation states
359: in the above analysis can be replaced with an HCO$^+$ transition. Both 18cm
360: main lines and a number of HCO$^+$ transitions have already been detected in
361: four molecular absorbers between $z \sim 0.25$ and $z \sim 0.9$
362: \citep{wiklind95,wiklind96a,wiklind96b,wiklind97,chengalur99,kanekar02,kanekar03}; the
363: detection of a single 6cm ``main'' line in one of these absorbers (which seems
364: the most promising of the remaining OH main lines) would thus be sufficient to carry
365: out the above measurement. We end this discussion by setting out the two simultaneous
366: equations obtained by using the 18cm and 6cm main lines and an HCO$^+$ line
367: in equation~(\ref{eqn:redshift}). Combining the sum of 18cm main line frequencies
368: with the HCO$^+$ frequency gives
369: \beq
370: \label{eqn:18cm-HCO}
371: \frac {\Delta z_{13}}{ 1 + {\bar z_{13}} }\;\; =\;\; 1.571 \frac{\Delta y}{y} -
372: 1.141 \frac{\Delta \alpha}{\alpha} \;\; .
373: \eeq
374:
375: \noi Similarly, combining the frequencies of the OH 6cm main line with an HCO$^+$
376: transition yields
377: \beq
378: \label{eqn:6cm-HCO}
379: \frac {\Delta z_{23}}{ 1 + {\bar z_{23}} }\;\; =\;\; -0.491\frac{\Delta y}{y} +
380: 2.983 \frac{\Delta \alpha}{\alpha} \;\; ,
381: \eeq
382:
383: \noi where $\Delta z_{ij} = z_{j} - z_{i}$, ${\bar z}_{ij} = \lb z_i + z_j \rb/2$
384: and the subscripts 1, 2 and 3 denote
385: the 18cm, 6cm and HCO$^+$ transitions respectively. A detection of the main 6cm OH
386: lines in any of the four cosmologically distant absorbers can thus be immediately
387: used in the above equations to measure $\dal$ and $\Delta y/y$.
388:
389: It has been emphasized in the introduction that the primary drawback of our
390: earlier analysis, using the 18cm OH lines (Paper~I), is that the frequency
391: separation between the main 18cm lines is only $\sim 1.957/(1 + z )$~MHz, where
392: $z$ is the absorption redshift. This dominated the errors in redshift measurements
393: and hence resulted in the large final errors in the analysis of the absorption lines
394: towards B0218+357. Of course, the advantage of this method is that it simultaneously
395: allows a measurement of changes in three fundamental parameters, $g_p$, $\alpha$ and
396: $y \equiv m_e/m_p$. In the present case, while we have a far higher accuracy in the
397: measurement (as the frequency sum of the 18cm main lines is larger than their frequency
398: difference by a factor of $\sim 1600$ and by an even higher factor for the
399: higher order transitions), the technique only allows an estimate of changes in $\alpha$
400: and $y$. However, these estimates can be replaced in the equation for the separation
401: between the satellite 18cm OH lines (equation~(11) in Paper~I) to constrain the
402: evolution of the proton g-factor $g_p$. While this last measurement would be a
403: factor of $\sim 30$ less sensitive than the measurements of $\dal$ and $\Delta y/ y$,
404: it is still more than 50 times more sensitive than estimates of $\Delta g_p/ g_p$
405: obtained using the earlier method (Paper~I); this is due to the fact that the
406: frequency separation between the 18cm satellite lines is $\sim 108/(1+z)$~MHz, fifty times
407: larger than that between the 18cm main lines.
408:
409: We note, in passing, that other ``Lambda-doubled'' systems are known to exist in the
410: laboratory and any of these could, in principle, be used in place of OH in a similar
411: calculation. However, to the best of our knowledge, multiple transitions have not been
412: detected in these other systems in astrophysical sources; we suspect that the strength
413: of cm-wave OH lines is likely to make OH the best candidate for such analyses. Further,
414: the present calculation is based on a perturbative treatment of the OH levels
415: \citep{vanvleck29, townes55}; more recent analyses (e.g.~\citealt{brown79}) use the
416: ``effective Hamiltonian'' approach, resulting in higher order effects. As pointed
417: out in Paper~I, these are unlikely to significantly affect our results.
418:
419: Finally, we estimate the accuracy that could be obtained in a measurement of
420: $\Delta \alpha / \alpha$ by the present method, for the $z \sim 0.885$ absorber
421: towards PKS~1830$-$21; the latter has the highest redshift of all known
422: 18cm OH absorbers \citep{kanekar02}. We simplify the analysis by assuming that $y$ is constant, so that
423: main lines of only two rotational states need be used; we will consider the
424: $\mathbf{^2\Pi_{3/2}, J = 3/2}$ and $\mathbf{^2\Pi_{1/2}, J = 1/2}$ states, i.e.
425: main lines at observing frequencies of $\sim 885$~MHz and $\sim 2520$~MHz,
426: respectively, for an absorber at $z = 0.885$. Next, a resolution of 2~kHz is
427: not unreasonable for radio spectroscopy with present-generation telescopes. Even
428: if we assume that the line centroids are only determined to this accuracy (i.e. that
429: sub-channel resolution is not obtained, via fitting to the line profile), the errors
430: on the redshifts of the sum of main line frequencies would be $\Delta z = 1 \times 10^{-6}$
431: and $\Delta z = 4 \times 10^{-7}$, for the 18cm and 6cm lines, respectively. Combining
432: equations~(\ref{eqn:18cm}) and (\ref{eqn:6cm}) in equation~(\ref{eqn:redshift}) (and assuming
433: $y$ to be constant) then gives $\Delta \alpha / \alpha = 1.5 \times 10^{-7}$,
434: the $1 \sigma$ accuracy in a measurement of changes in $\alpha$ from $z \sim 0.885$
435: to today. This is significantly better than the precision obtained in the best optical
436: studies today (e.g. $\Delta \alpha / \alpha = 6 \times 10^{-7}$; \citet{humchand04}).
437: Note, however, that the above errors do not take into account systematic effects, i.e.
438: the possibility of relative motions between the 18cm and 6cm absorbing clouds; as
439: mentioned earlier, observations of lines from other OH rotational states would help
440: to constrain such systematics.
441:
442: In summary, we have demonstrated a new technique to simultaneously measure any
443: evolution in two fundamental constants $\alpha$ and $y \equiv m_e/m_p$, using OH
444: ``main'' absorption lines. The method is $\sim$ three orders of magnitude more
445: sensitive than that described by Chengalur \& Kanekar
446: (2003), which utilised the four OH 18cm lines for the analysis. The increase in
447: sensitivity comes from the use of OH ``main'' lines arising from different OH
448: rotational states, rather than the difference between the frequencies of lines
449: arising from the same state. The technique requires the detection of main lines
450: from three OH rotational states or, alternately, two OH states and one HCO$^+$
451: transition, in the same absorber. The large number of OH rotational states also
452: allows a simple consistency check of any measurement by a search for the main
453: lines of one further OH state. Both 18cm main lines and a number of HCO$^+$
454: transitions have been detected in absorption in four extra-galactic molecular
455: absorbers. The detection of a single 6cm main line in any of these systems would
456: thus be sufficient to simultaneously measure (or constrain) changes in $\alpha$
457: and $y \equiv m_e/m_p$.
458:
459: \section{Acknowledgments}
460: We are grateful to Rajaram Nityananda for very useful discussions
461: on the energy levels of the OH ground state.
462:
463:
464: \bibliographystyle{mn2e}
465: \bibliography{ms}
466:
467: \end{document}
468: