astro-ph0310764/ms.tex
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[prl,showpacs,amsmath,amssymb]{revtex4}
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{graphicx}% Include figure files
31: \usepackage{dcolumn}% Align table columns on decimal point
32: \usepackage{bm}% bold math
33: \usepackage{natbib}
34: 
35: %\nofiles
36: 
37: \newcommand{\al}{\ensuremath{\alpha}}
38: \newcommand{\dal}{\ensuremath{\Delta \alpha/ \alpha}}
39: \newcommand{\gp}{\ensuremath{{g}_{p}}}
40: \newcommand{\lqcd}{\ensuremath{\Lambda_{\rm QCD}}}
41: \newcommand{\y}{\ensuremath{{m}_{e}/{m}_{p}}}
42: \newcommand{\dy}{\ensuremath{\Delta {m_e}/{m_p}/{m_e}/{m_e}}}
43: 
44: 
45: \newcommand{\beq}{\begin{equation}}
46: \newcommand{\eeq}{\end{equation}}
47: \newcommand{\noi}{\noindent}
48: \newcommand{\lb}{\left(}
49: \newcommand{\rb}{\right)}
50: \newcommand{\lsb}{\left[}
51: \newcommand{\rsb}{\right]}
52: 
53: \begin{document}
54: 
55: %\preprint{APS/123-QED}
56: 
57: \title{Constraining the variation of fundamental constants using 18cm OH lines}
58: \author{Jayaram N. Chengalur}
59: \email{chengalur@ncra.tifr.res.in}
60: \affiliation{ NCRA/TIFR, P. O. Bag 3, Ganeshkhind, Pune 411007, India}
61: %\altaffiliation[Also at ]{Physics Department, XYZ University.}
62: \author{Nissim Kanekar}
63: \email{nissim@astro.rug.nl}
64: \affiliation{ Kapteyn Institute, Groningen University, The Netherlands}
65: 
66: \date{\today}% It is always \today, today,
67:              %  but any date may be explicitly specified
68: 
69: \begin{abstract}
70: We describe a new technique to estimate variations in the fundamental 
71: constants using 18cm OH absorption lines. This has the advantage that all lines 
72: arise in the same species, allowing a clean comparison between the measured 
73: redshifts. In conjunction with one additional transition (for example, an 
74: HCO$^+$ line), it is possible to simultaneously measure changes in $\alpha$, 
75: $g_p$ and $y \equiv m_e/m_p$. At present, only the 1665~MHz and 1667~MHz lines 
76: have been detected at cosmological distances; we use these line redshifts in 
77: conjunction with those of HI 21cm and mm-wave molecular absorption in a 
78: gravitational lens at $z\sim 0.68$ to constrain changes in the above three 
79: parameters over the redshift range $0 < z \lesssim 0.68$. While the 
80: constraints are relatively weak ($\lesssim$ 1 part in $10^3$), this is the 
81: first simultaneous constraint on the variation of all three parameters. 
82: We also demonstrate that either one (or more) of $\alpha$, $g_p$ and 
83: $y$ must vary with cosmological time or there must be systematic velocity 
84: offsets  between the OH, HCO$^+$ and HI absorbing clouds.
85: \end{abstract}
86: 
87: \pacs{98.80.-k,98.80.Es,98.58.-w,33.15.Pw}% PACS, the Physics and Astronomy
88:                              % Classification Scheme.
89: %\keywords{Suggested keywords}%Use showkeys class option if keyword
90:                               %display desired
91: \maketitle
92: 
93: \section{\label{sec:intro} Introduction}
94: 
95: 
96: The recent claim by Webb et al. \cite{webb99,webb01} that the fine structure constant 
97: $\alpha$ evolves with redshift, with $\dal = (-1.88 \pm 0.53) \times 10^{-5}$ from 
98: $z \sim 1.6$ to today and $\dal = (-0.72 \pm 0.18) \times 10^{-5}$ for $0.5 < z < 3.5$~
99: (but see \cite{bekenstein03}) has spurred interest in the possibility that the numerical 
100: values of the fundamental constants change with time. Theories that can account for 
101: such variations include extra-dimensional Kaluza-Klein theories and superstring 
102: theories. In such models, the values of the coupling constants depend on the 
103: expectation values of some cosmological scalar field(s); changes in the values 
104: of the coupling constants are thus to be expected if this field varies 
105: with location and time. Further, depending on the details of the theory, all of 
106: the different coupling constants (such as $\alpha$, the proton g-factor $g_p$, 
107: the electron-proton mass ratio $y \equiv m_e/m_p$, the gravitational constant $G$, etc)
108: could, in principle, vary simultaneously.
109: For example, Calmet~\&~Fritzsch \cite{calmet02} and Langacker~et al. 
110: \cite{langacker02} find that variations in the value of $\alpha$ should be 
111: accompanied by much larger changes (by $\sim 2$ orders of magnitude) in the 
112: value of $y$. However, Ivanchik~et al.  \cite{ivanchik03} constrain the variation 
113: in $y$ to be $(3.0 \pm 2.4) \times 10^{-5}$ over the redshift range  
114: $ 0< z <3$, comparable to the change claimed in the fine structure constant.
115: A review of the available experimental and observational measurements 
116: on the variability of the coupling constants can be found in \cite{uzan03}.
117: 
118: One of the main problems in most of the astrophysical techniques used to measure (or 
119: constrain) the values of the different constants (e.g. \cite{carilli00,webb01,ivanchik03})
120: is that they involve a comparison between the redshifts of spectral lines of
121: different species (e.g. the HI 21cm line, millimetre-wave molecular lines 
122: and optical fine structure lines \cite{webb01,carilli00}). These species are 
123: unlikely to all arise at the same physical location in a gas cloud and might thus 
124: have systematic velocity offsets relative to each other; the redshift differences 
125: may thus be dominated by these effects rather than the measurement errors
126: (i.e. the spectral resolution, which can be quite small, $\sim$~few km/s, for 
127: HI 21cm and mm-wave molecular absorption spectra). Conclusions drawn from a 
128: comparison between different species might thus well be incorrect.
129: 
130: Clearly, the best way to test the variation of the coupling constants is to use 
131: lines originating from a {\it single species}, but with different 
132: dependences on these constants. Further, since many constants may be varying 
133: simultaneously, it would be very useful if one could simultaneously measure the 
134: changes in a number of constants from this single species, rather than assuming 
135: that changes occur in only one of the constants and that the others remain unchanged.
136: We present here a new technique which satisfies both these requirements, using 
137: the 18cm lines of the OH radical.
138: 
139: 	The ground $^2\Pi_{3/2}$ J=3/2 state of OH is split into two 
140: levels by $\Lambda$-doubling and each of these $\Lambda$-doubled levels 
141: further split into two hyperfine states. Transitions between these levels 
142: lead to four spectral lines with wavelength $\sim 18$cm. Transitions with 
143: $\Delta F=0$ are called the ``main'' lines, arising at rest frequencies 
144: of 1665.4018~MHz and 1667.3590~MHz, while transitions with $\Delta F = 1$ 
145: are called ``satellite'' lines, with rest frequencies of 1612.2310~MHz and 
146: 1720.5299~MHz. Since the four OH lines arise from two very different physical 
147: processes, viz. $\Lambda$-doubling and hyperfine splitting, the transition 
148: frequencies have different dependences on the 
149: fundamental constants. A perturbative treatment of the OH molecule has been 
150: carried out by Dousmanis et al. \cite{dousmanis55} (see also \cite{townes55});
151: we use the expressions derived in these references to determine the dependence
152: of linear combinations of the four OH line frequencies on $\alpha$, $g_p$ and 
153: $y \equiv m_e/m_p$ and show that it is possible to simultaneously measure 
154: changes in both $\alpha$ and $y$, if all four line frequencies are known 
155: (assuming that $g_p$ does not vary with time). Since OH and HCO$^+$ column densities
156: are observed to be tightly correlated, both in the Galaxy \cite{liszt96} and out to $z
157: \sim 1$ \cite{kanekar02}, these species are likely to arise at the same physical 
158: location; a comparison between the redshifts of the four 18cm OH lines and 
159: the HCO$^+$ line should thus allow one to constrain the evolution of all the 
160: above three parameters. We use our observations of the OH main lines 
161: in the $z \sim 0.6846$ gravitational lens towards B0218+357, in tandem with 
162: published HI 21cm and mm-wave molecular redshifts, to constrain $\Delta y/y$ 
163: between $z \sim 0.68$ and today. Finally, as this work was being written up, 
164: an analysis on the use of OH lines to constrain changes in fundamental constants 
165: was also carried out by Darling \cite{darling03}; the latter, however, only 
166: considers variations in the fine structure constant $\alpha$.
167: 
168: \section{ Constraints from Radio Spectral Lines}
169: 
170: Consider two transitions whose rest frame frequencies $\nu_i(z), i = 1, 2$ 
171: depend on redshift, due to the evolution of various fundamental constants, 
172: such as $\al, \gp, \y$, etc. If the lines arise in a source at a ``true'' 
173: redshift $z$, the {\it measured} redshift $\hat z_i$ of each line is given by
174: 
175: \beq
176: \lb 1 + {\hat z_i} \rb^{-1} \;\;=\;\; \frac{\nu_i(z) }{ \lb 1 + z \rb \nu_i(0) }
177: \;\;=\;\; \frac{1 + \lsb \Delta \nu_i(z)/\nu_i(0) \rsb }{1 +z} \;\; ,
178: \eeq
179: \noi where $\Delta \nu_i (z) \equiv \nu_i(z) - \nu_i(0)$; note that $\Delta 
180: \nu_i (z) = 0$ if the fundamental constants do not change with time. The 
181: first order difference between the two measured redshifts $\Delta z = 
182: \hat{z}_1 -\hat{z}_2$ is then 
183: 
184: \beq 
185: \label{eqn:redshift}
186: \frac {\Delta z }{ 1 + {\bar z} }\;\; =\;\; \lsb \frac{ \Delta \nu_2}{\nu_2(0)} \rsb - 
187: \lsb \frac{\Delta \nu_1}{\nu_1(0)} \rsb \;\; ,
188: \eeq
189: 
190: \noi where $\bar z$ is the mean measured redshift. Given two spectral lines (or 
191: linear combinations of line frequencies) with different dependences on some 
192: fundamental parameter, one can thus use the differences between the measured 
193: redshifts to constrain the evolution of the parameter in question. 
194: 
195: In the case of the four 18cm OH lines, the following three independent relations 
196: have been shown to be satisfied by the line frequencies \cite{townes55,dousmanis55}; 
197: note that the lines must also satisfy the constraint $\nu_{1665} + \nu_{1667} 
198: = \nu_{1720} + \nu_{1612}$ 
199: 
200: \beq
201: \label{eqn:sum1}
202: \nu_A\;\; \equiv\;\; \nu_{1667} + \nu_{1665}\;\; =\;\; q_\Lambda\lsb 
203: \lb 2 + \frac{A'}{B'} \rb \lb 1 - \frac{2 - A/B}{X} \rb -
204: \frac{12}{X} \rsb \;\; ,
205: \eeq
206: \beq
207: \label{eqn:diff1}
208: \nu_B\;\; \equiv\;\; \nu_{1667} - \nu_{1665}\;\; =\;\; \frac{8d \lb X - 2 + A/B \rb }{15 X}  \: \: \:\:\:\: \mathrm{and}
209: \eeq
210: \beq
211: \label{eqn:diff2}
212: \nu_C\;\; \equiv\;\; \nu_{1720} - \nu_{1612}\;\; =\;\; 
213: \frac{4}{15X} \lsb 2a \lb 2X + 2 - A/B \rb  +  12b +\lb b+c \rb 
214: \lb X + 4 - 2A/B \rb \rsb
215: \eeq
216: 
217: \noi Equations (\ref{eqn:sum1}), (\ref{eqn:diff1}) \& (\ref{eqn:diff2}) correspond to the 
218: energy split due to $\Lambda$~doubling and the difference and sum of the hyperfine splits 
219: in the two $\Lambda$~doubled levels respectively. Here, $X \equiv \lsb \lb 
220: A/B \rb \{ {\lb A/B\rb} - 4 \}+ 16 \rsb^{1/2}$, 
221: $A$ is the fine structure interaction constant, $B$, the rotational constant, 
222: $A'$ and $B'$, the off-diagonal matrix elements of these operators, 
223: $q_\Lambda \approx 4B^2/h\nu_e$ (where $h\nu_e$ is the energy difference between 
224: the ground and first excited electronic state), and, finally, $a$, $b$, $c$ and $d$ 
225: are ``hyperfine constants'' \cite{dousmanis55}, whose experimental values are 
226: $a = 86.012 \pm 0.002$~MHz, $ b = -116.719 \pm 0.008$~MHz, $c   = 130.75 \pm 0.01$~MHz 
227: and $d = 56.632 \pm 0.004$~MHz \cite{coxon79}. Numerically, $A/B = -7.547$ and 
228: $A'/B' = -6.073$ \cite{townes55}. These quantities have the following dependences on 
229: the fundamental constants $\alpha$, $y$, $R_\infty \equiv m_e e^4/\hbar ^3 c$ and $g_p$ : 
230: $A' \propto A \propto \alpha^2 R_\infty \;\;$, $B' \propto B \propto y R_\infty \;\;$ 
231: and $a, b, c, d \propto g_p \alpha^2 y R_\infty$ \cite{townes55}. For the rotational constant $B$, 
232: we have assumed, as usual (e.g. \cite{murphy01}), that variations in $(m_p/M)$, 
233: which are suppressed by a factor $m_p/U \sim 100$ (where $M$ is the reduced mass and 
234: U the binding energy) can be ignored. Thus, we have 
235: $ \lsb A'/B' \rsb \propto \lsb A/B \rsb \propto \lb \alpha^2/y \rb $.
236: Replacing the above scalings in equation~(\ref{eqn:sum1}) for $\nu_A$, we obtain
237: $\nu_A \propto y^2 R_\infty F\lb \alpha^2/y\rb \;\;$, where $F \equiv F \lb \beta \rb $ 
238: is a function which depends only on the ratio $\beta \equiv A/B \propto \alpha^2/y$ 
239: and is defined by 
240: \beq
241: F\lb \beta \rb = \lsb \lb 2 + \frac{6.073}{7.547}\beta \rb \lb 1 + 
242: \frac{2 - \beta}{X\lb\beta\rb} \rb + \frac{12}{X\lb \beta\rb} \rsb
243: \eeq
244: \noi Thus,
245: \begin{eqnarray}
246: \frac{\Delta \nu_A }{\nu_A} &=& 2\frac{\Delta y}{y} + \frac{\Delta R_\infty}{R_\infty}
247: + \frac{\Delta F \lb \beta \rb} {F\lb\beta \rb}\\
248: &=& 2\frac{\Delta y}{y} + \frac{\Delta R_\infty}{R_\infty} 
249: + \frac{\beta}{F} \frac{dF}{d\beta} \lsb 2 \frac{\Delta \alpha}{\alpha} - 
250: \frac{\Delta y}{y} \rsb 
251: \end{eqnarray}
252: 
253: \noi Evaluating the quantity on the right-hand-side of the above equation, we 
254: obtain 
255: \beq 
256: \label{eqn:temp1}
257: \frac{\Delta \nu_A }{\nu_A} = 2.571 \frac{\Delta y}{y} - 
258: 1.141 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
259: \eeq
260: 
261: \noi In similar fashion, equations~(\ref{eqn:diff1}) and (\ref{eqn:diff2}) yield 
262: \beq 
263: \label{eqn:temp2}
264: \frac{\Delta \nu_B }{\nu_B} = 2.442\frac{\Delta y}{y} -
265: 0.883 \frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
266: + \frac{\Delta g_p}{g_p} \:\:\:\:\:\: \mathrm{and}
267: \eeq
268: \beq
269: \label{eqn:temp3}
270: \frac{\Delta \nu_C }{\nu_C} = 0.722\frac{\Delta y}{y} +
271: 2.557\frac{\Delta \alpha}{\alpha} + \frac{\Delta R_\infty}{R_\infty}
272: + \frac{\Delta g_p}{g_p}
273: \eeq
274: 
275: \noi Equations~(\ref{eqn:temp1} -- \ref{eqn:temp3}) govern the way in which 
276: a change in one of the fundamental constants affects the line rest 
277: frequencies. Combining them in pairs in equation~(\ref{eqn:redshift}) yields :
278: 
279: \beq
280: \label{eqn:z12}
281: \frac {\Delta z_{AB} }{ 1 + {\bar z_{AB}} } = 
282: \lsb \frac{\Delta \nu_B}{\nu_B} - \frac{\Delta \nu_A}{\nu_A} \rsb 
283: = -0.129 \frac{\Delta y}{y} + 0.258\frac{\Delta \alpha}{\alpha} 
284: + \frac{\Delta g_p}{g_p} \;\; ,
285: \eeq
286: 
287: \beq
288: \label{eqn:z13}
289: \frac {\Delta z_{AC} }{ 1 + {\bar z_{AC}} } = 
290: \lsb \frac{\Delta \nu_C}{\nu_C} - \frac{\Delta \nu_A}{\nu_A} \rsb 
291: = -1.849\frac{\Delta y}{y} 
292: + 3.698\frac{\Delta \alpha}{\alpha} + \frac{\Delta g_p}{g_p} \:\:\:\:\:\: \mathrm{and}
293: \eeq
294: 
295: \noi If all four OH lines are detected in absorption in a single cosmological 
296: system, we thus have two independent equations relating the differences in measured 
297: redshifts to the changes in $\alpha$, $g_p$ and $y \equiv (m_e/m_p)$. Most 
298: analyses (using other lines) assume that $g_p$ remains unchanged and then estimate
299: the variation of $\alpha$ \cite{murphy01,carilli00}. If one makes the same assumption
300: in the case of the OH lines,
301: equations~(\ref{eqn:z12} -- \ref{eqn:z13}) immediately allow one to simultaneously 
302: solve for changes in both $\alpha$ and $(m_e/m_p)$. Of course, one further 
303: equation is needed to simultaneously constrain the evolution of all three constants.
304: Candidates include the HI 21cm and mm-wave molecular lines. The best of these 
305: are likely to be the HCO$^+$ lines since HCO$^+$ and OH column densities are 
306: found to show a strong correlation (extending over more than two orders of magnitude
307: in column density) both in the Galaxy \cite{liszt96} and out to $z \sim 1$; 
308: \cite{kanekar02}, this suggests that HCO$^+$ and OH are located in the same 
309: region of the molecular cloud. Since the HCO$^+$ line arises from a rotational 
310: transition, we have
311: \beq
312: \label{eqn:mm}
313: \frac{\Delta \nu_{HCO^+}}{\nu_{HCO^+}} = \frac{\Delta y}{y}
314: + \frac{\Delta R_\infty}{R_\infty} \;\; ,
315: \eeq
316: 
317: Equations~(\ref{eqn:z12}), (\ref{eqn:z13}) and (\ref{eqn:mm}) 
318: all have the same dependence on the Rydberg constant $R_\infty$, which hence 
319: cancels out (note that $R_\infty$ itself depends on $\alpha$ 
320: through the relation $R_\infty \equiv m_e e^4/\hbar ^3 c$). One might thus 
321: combine HCO$^+$ absorption redshifts with those derived from the 18cm OH 
322: lines to provide the last equation needed to solve for the evolution of 
323: $\alpha$, $y$ and $g_p$. This would allow a simultaneous measurement 
324: of all three constants, which has been hitherto impossible. We note that 
325: it is also possible to use other OH $\Lambda$-doubled transitions to 
326: simultaneously constrain the variation of the above parameters; this 
327: is discussed elsewhere \cite{kanekar04}, for the ``main'' OH lines. It 
328: would also be interesting to carry out 
329: an analysis similar to that of Bekenstein \cite{bekenstein03} to test 
330: whether these analyses of the OH lines are also affected by the possibility 
331: that the Hamiltonians involved might vary with time (i.e. {\it dynamical} 
332: variability of the different parameters); this is beyond the scope of the 
333: present paper. It should also be pointed out that the present calculation is 
334: based on an analysis of the 
335: OH levels using perturbation theory \cite{dousmanis55}; more recent 
336: analyses \cite{brown79} use the ``effective Hamiltonian'' approach, giving
337: rise to higher order effects. While it would be interesting to attempt 
338: the present calculation in the latter framework, we emphasise that it 
339: should only result in small changes to the coefficients in 
340: equations~(\ref{eqn:temp1} -- \ref{eqn:temp3}) and does not affect the 
341: validity of our approach. Finally, while other ``Lambda-doubled'' systems 
342: could, in principle, be used for a similar analysis, none of these have 
343: multiple transitions detected in astrophysical objects (to the best of 
344: our knowledge). While it would be interesting to carry out searches for 
345: these other transitions, we suspect that OH is yet likely to prove the 
346: best candidate because of the strength of its multiple lines.
347: 
348: 
349: \section{Application to the $z = 0.6846$ absorber towards B0218+357}
350: \label{sec:0218}
351: 
352: While the above analysis shows that one can use the four 18cm OH lines 
353: (in conjunction with the HCO$^+$ transition) to constrain the evolution 
354: of three separate fundamental parameters, the weaker, satellite 1612~MHz 
355: and 1720~MHz lines have so far not been detected at cosmological distances. 
356: Observations are currently being scheduled to carry out deep searches for these 
357: lines in the four known OH absorbers at intermediate redshift 
358: \cite{chengalur99,kanekar02,kanekar03}. For the present, we will instead use the 
359: detected 1665~MHz and 1667~MHz transitions (i.e. equations~(\ref{eqn:temp1}) 
360: and (\ref{eqn:temp2})) along with the HI 21cm and millimetre (HCO$^+$) lines 
361: in the $z = 0.6846$ absorber towards B0218+357 to estimate changes in the 
362: fundamental parameters (assuming that the lines arise in the same gas cloud). 
363: Since the HI 21cm frequency arises from a hyperfine split and is hence 
364: proportional to $g_p y \alpha^2 R_\infty$, we have
365: \beq
366: \label{eqn:HI}
367: \frac{\Delta \nu_{21}}{\nu_{21}} = \frac{\Delta y}{y}
368: + 2 \frac{\Delta \alpha}{\alpha} + \frac{\Delta g_p}{g_p} 
369: + \frac{\Delta R_\infty}{R_\infty}
370: \eeq
371: 
372: \noi Equations~(\ref{eqn:temp1}), (\ref{eqn:temp2}), (\ref{eqn:mm}) and 
373: (\ref{eqn:HI}) can now be solved to measure changes in $\alpha$, $g_p$ and 
374: $m_e/m_p$. The HI 21cm redshift is $z_{HI} = 0.684676 \pm 0.000005$ 
375: \cite{carilli00}, while that of the HCO$^+$ absorption lines is 
376: $z_{HCO^+} = 0.684693 \pm 0.000001$ \cite{wiklind97}. Our new GMRT OH absorption 
377: spectra towards B0218+357 \cite{kanekar03} yield the following redshifts 
378: for the sum and difference of the 1665~MHz and 1667~MHz line frequencies 
379: : $z_{sum} = 0.684682 \pm 0.0000056$ and $z_{diff} = 0.685780 \pm 
380: 0.0067$. A simultaneous solution of equations~(\ref{eqn:temp1}), 
381: (\ref{eqn:temp2}), (\ref{eqn:mm}) and (\ref{eqn:HI}) then yields 
382: $(\Delta \alpha / \alpha) = (-0.38 \pm 2.2) \times 10^{-3}$, 
383: $(\Delta y/ y) = (-0.27 \pm 1.6) \times 10^{-3}$ and $(\Delta g_p/ g_p) 
384: = (-0.77 \pm 4.2) \times 10^{-3}$. Since the error on $z_{diff}$ is far higher 
385: than the other errors, this dominates the errors on the above estimates and results in
386: relatively uninteresting upper limits on changes in the three constants. We emphasise,
387: however, that, to the best of our knowledge, this is the first time that a simultaneous
388: constraint on the variation of these three fundamental parameters has 
389: been obtained in a cosmologically distant object. Further, the 
390: weakness of the constraint arises from the relatively small difference between 
391: the frequencies of the main lines; this is clearly not a fundamental limitation 
392: but depends entirely on the sensitivity of the observations.
393: 
394: Next, if we {\it assume} that $y \equiv m_e/m_p$ is constant, 
395: equations~(\ref{eqn:temp1}) and (\ref{eqn:mm}) yield $\lsb \dal \rsb 
396: = - 5.7 \pm 3.0 \times 10^{-6}$. This is more than $3\sigma$ deviant from
397: the estimate $\lsb 2\Delta \alpha/\alpha \rsb = 1 \pm 0.3 \times 10^{-5}$ of
398: Carilli et al. (2000) \cite{carilli00}. Since the latter analysis assumed
399: that $g_p$ was constant, the difference between the two estimates implies
400: either that the assumptions that $y$ and/or $g_p$ are constant is
401: unjustified or that systematic velocity offsets do exist between the three
402: species.
403: 
404: Further progress can be made if one of the three parameters, $\alpha$, $g_p$ 
405: and $y$ is assumed to not change with time (while retaining the assumption that 
406: velocity offsets are not significant). We can then avoid having to use the 
407: equation governing the difference between the OH redshifts and can thus obtain a 
408: far stronger limit on changes in the remaining two quantities. For example, if 
409: we assume (as is often done, e.g. \cite{carilli00}) that $g_p$ remains constant, 
410: the HI 21cm and mm-wave molecular line redshifts imply $\lsb \dal  \rsb = (5 \pm 1.5)\times 10^{-6}$ 
411: \cite{carilli00}. Combining equations~(\ref{eqn:temp1}) and (\ref{eqn:mm}) 
412: then yields $\lsb \Delta y / y\rsb = (7.8\pm 2.4) \times 10^{-6} $.
413: Similarly, if we assume that $y \equiv m_e/m_p$ is constant, 
414: equations~(\ref{eqn:temp1}) and (\ref{eqn:mm}) yield $\lsb \dal \rsb = 
415: (- 5.7 \pm 3.0) \times 10^{-6}$ and $\lsb \Delta g_p /g_p\rsb = (2.2 \pm 
416: 0.67) \times 10^{-6}$. Finally, if $\alpha$ remains unchanged, we obtain 
417: $\lsb \Delta y/y \rsb = (4.2 \pm 2.2) \times 10^{-6}$ and $\lsb \Delta g_p/g_p 
418: \rsb = (1 \pm 0.3) \times 10^{-5}$. The above limits on $\Delta y/y$ are a 
419: factor of $\sim 4$ stronger than the best earlier limits on changes in this 
420: quantity \cite{ivanchik03}. It is very interesting that all three cases 
421: result in a higher than $3\sigma$ significance for the variation 
422: of at least one of the parameters. This implies either that one (or more) 
423: of these parameters indeed varies with cosmological time or that systematic
424: motions between the three species cause the above uncertainties (which only 
425: include measurement errors) to be under-estimated.
426: 
427: Finally, we note that, while the OH 1667~MHz and 1665~MHz, HCO$^+$ and 
428: HI lines have been detected in four absorbers at intermediate redshifts, two 
429: of the absorbers (PKS1413+135 and B2~1504+377) are believed to have velocity 
430: offsets between the HI and HCO$^+$ redshifts \cite{carilli00,wiklind96a}. 
431: The last system, at $z \sim 0.889$ towards PKS1830$-$21, does not at 
432: present have OH data of sufficiently high quality to carry out the above 
433: analysis.
434: 
435: In summary, we have demonstrated a new technique to simultaneously measure 
436: the evolution of the three fundamental constants $\alpha$, $g_p$ and $m_e/m_p$, 
437: using 18cm OH absorption lines in conjunction with one additional transition, 
438: (which could be an HCO$^+$ mm-wave line). At present, only the 1665~MHz and 1667~MHz 
439: ``main'' OH lines have been discovered at cosmological distances; we have 
440: used these line redshifts in conjunction with those of HI 21cm absorption and 
441: millimetre-wave molecular lines to constrain the variation of $y$, $g_p$ and 
442: $\alpha$ between $z = 0.6846$ and today.  We argue that one (or more) of 
443: the parameters $\alpha$, $y$ and $g_p$ must vary with cosmological time, 
444: unless systematic velocity offsets exist between the above three species. 
445: The constraints placed on changes in the parameter $y \equiv m_e/m_p$ 
446: (assuming that either $\alpha$ or $g_p$ are constant) are a factor of 
447: $\sim 4$ stronger than earlier limits on variations in this parameter.
448: 
449: \begin{acknowledgments}
450: 	We are grateful to Rajaram Nityananda for very useful discussions
451: on the energy levels of the OH ground state.
452: \end{acknowledgments}
453: 
454: 
455: \bibliography{ms}
456: 
457: \end{document}
458: