A stringent limit_on_a_drifting_photon_to_electron_mass_ratio_from_alcohol_in_the_earley_universe
A Stringent Limit on a Drifting Proton-to-Electron Mass Ratio from Alcohol in the Early Universe Julija Bagdonaite et al. Science 339, 46 (2013); DOI: 10.1126/science.1224898 This copy is for your personal, non-commercial use only. If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Downloaded from www.sciencemag.org on January 4, 2013 Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here. The following resources related to this article are available online at www.sciencemag.org (this information is current as of January 3, 2013 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/339/6115/46.full.html Supporting Online Material can be found at: http://www.sciencemag.org/content/suppl/2012/12/12/science.1224898.DC1.html This article cites 30 articles, 3 of which can be accessed free: http://www.sciencemag.org/content/339/6115/46.full.html#ref-list-1Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright2013 by the American Association for the Advancement of Science; all rights reserved. The title Science is aregistered trademark of AAAS.
REPORTS to radio astronomy. Here we use the extreme sensitivity of methanol (CH3OH) (14, 15) to probe A Stringent Limit on a Drifting the variation of the proton-to-electron mass ratio m over cosmic time. Methanol (Fig. 1A) is the simplest alcohol Proton-to-Electron Mass Ratio and consists of a hydroxyl group attached to a methyl group. The C-O bond is flexible, allow- from Alcohol in the Early Universe ing the hydroxyl group to rotate with respect to the methyl group. This so-called internal rotation is strongly hindered by the repulsion between the Julija Bagdonaite,1 Paul Jansen,1 Christian Henkel,2,3 Hendrick L. Bethlem,1 hydrogen atoms of the different groups, result- Karl M. Menten,2 Wim Ubachs1* ing in a threefold barrier (Fig. 1B). If the barrier were infinitely high, the levels in the torsional The standard model of physics is built on the fundamental constants of nature, but it does not well would be degenerate. Quantum mechanical provide an explanation for their values, nor require their constancy over space and time. tunneling through the barriers lifts this degen- Here we set a limit on a possible cosmological variation of the proton-to-electron mass ratio m eracy, resulting in three levels that are labeled by comparing transitions in methanol observed in the early universe with those measured in according to symmetry: A, E1, and E2 (16). Be- Downloaded from www.sciencemag.org on January 4, 2013 the laboratory. From radio-astronomical observations of PKS1830-211, we deduced a constraint cause the symmetry of the nuclear wave function of ∆m/m = (0.0 T 1.0) × 10−7 at redshift z = 0.89, corresponding to a look-back time of 7 billion is preserved in radiative transitions as well as years. This is consistent with a null result. in (nonreactive) collisions, the A and E levels of methanol can be regarded as belonging to he standard model of particle physics, the observations with the world’s largest optical two separate chemical species T theory describing symmetries and forces of nature at the deepest level, does not provide an intrinsic explanation for the values of telescopes, have yielded a limit at the level of ∆m/m < 10−5 for look-back times of 12 billion years (8, 9). The sensitivity coefficient, Km, of a transition with frequency n is defined by ∆n/n = Km × ∆m/m. The frequencies of pure rotational transitions, such the fundamental coupling constants, nor does it Inversion transitions of ammonia (NH3) were as the transitions indicated by the red and orange prohibit that the fundamental constants depend found to be ~100 times more sensitive to m-variation arrow in Fig. 1C, are inversely proportional to on time and space. In contrast, Einstein’s equiv- than H2 transitions (10, 11). Astronomical obser- the reduced mass of methanol and hence to alence principle, a basic assumption of general vations of NH3, in the microwave or radio range the proton-to-electron mass ratio. Consequently, relativity, assumes that the laws of nature, and of the electromagnetic spectrum, led to stringent these have sensitivity coefficients equal to –1. hence the fundamental constants are independent 1s constraints at the level of (1.0 T 4.7) × 10−7 The frequencies of pure torsional transitions— of a local reference system. Some cosmological (12) and (–3.5 T 1.2) × 10−7 (13). This has shifted which are not allowed in methanol—depend ex- scenarios aimed at explaining the fine-tuning the paradigm for probing m-variation from optical ponentially on the reduced moments of inertia of between fundamental constants sketch an evolv- ing mechanism, where minimally varying con- Fig. 1. (A) Pictorial rep- C 145 J Ts J Ts stants are crucial for reaching the present state A resentation of the metha- 4 + A 1 E1 of complexity in the universe (1). Theoretical ap- nol molecule. (B) Potential ± 2 A proaches involving additional scalar fields have energy as a function of the 2 E 3 E2 Energy (cm-1) imposed bounds on varying constants through 140 ± dihedral angle between 1 A tests of the weak equivalence principle (2). In the the OH group and one of 1 +E 3 A past decade the search for small variations of the CH bonds in the methyl 0 E 2 E2 dimensionless fundamental constants over cosmo- group. V3 denotes the bar- 135 logical time scales has become an active exper- rier height. The horizontal + 1 E2 2 A imental endeavor, in particular because accurate lines represent the energy for the levels in the torsion- 130 |K| = 1 measurements of spectral lines of atoms at high + 1 A redshift have provided indication for a possible vibrational ground state, + 0 A variation of the fine structure constant a, either n t = 0, and first excited temporally (3, 4) or spatially (5, 6). state, n t = 1. (C) Energy B K=0 A second dimensionless fundamental constant level structure of the torsion- 400 νt = 1 Energy (cm-1) m, representing the proton-to-electron mass ratio rotation ground state of methanol. Each level is 300 mp/me, probes the cosmological evolution of the nuclear versus the electroweak sector in the stan- labeled according to its 200 torsional symmetry Ts, to- V3 dard model. A search for a possible drift of m has tal angular momentum J, 100 νt = 0 been made operational by comparing observa- and its projection K on 0 tions of spectral lines of the hydrogen molecule the molecule-fixed axis. (H2) in distant galaxies with accurate laboratory The energy-level structure 0 π/3 2π/3 π 4π/3 5π/3 2π measurements (7). These investigations, based on of methanol resembles that Torsional angle (rad) 1 of a prolate symmetric top, Department of Physics and Astronomy, VU University Amster- dam, De Boelelaan 1081, 1081 HV Amsterdam, Netherlands. with the difference that each K manifold is offset depending on its torsional symmetry. Levels of A-symmetry 2 Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, with |K| > 0 are split by the slight asymmetry of the molecule. Levels of E1 and E2 symmetry in the K = 0 53121 Bonn, Germany. 3Astronomy Department, King Abdulaziz manifold cannot be distinguished and are labeled as E. The four transitions observed in this study are University, Post Office Box 80203, Jeddah, Saudi Arabia. indicated by the four arrows. The transitions indicated by the red and orange arrows are pure rotational *To whom correspondence should be addressed. E-mail: transitions and have a sensitivity coefficient of –1. The transitions indicated by the blue and green arrows firstname.lastname@example.org are mixed torsion-rotation transitions and have sensitivity coefficients of –32.8 and –7.4, respectively.46 4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org
REPORTSthe methyl and hydroxyl groups and have a changed—is strongly enhanced. The sensitivity but because of a number of favorable prop-sensitivity coefficient of –2.5. The sensitivity coefficients for different transitions in methanol erties, the effect is exceptionally large in meth-of mixed transitions—i.e., transitions in which range from –42 to +53. These enhancements anol (17).both the internal and overall rotation state is occur generally in every internal rotor molecule, Methanol is abundantly present in the uni- verse, and more than a 1000 lines have been recorded in our galaxy (18). So far, searches forFig. 2. Methanol absorp- methanol absorption in far-distant galaxies have Residuals Kµ = -32.8 yielded detection only in the gravitational lenstion lines on a local standardof rest (LSR) velocity scale 0 Combined system PKS1830-211 (19). A limit on ∆m/m (20)relative to z = 0.88582, has been previously derived on the basis of twoobserved toward PKS1830- -0.1 methanol lines. We present a comprehensive 23-02-2012211. The transitions and study of radio astronomical observations of fourtheir approximate observed 25-02-2012 methanol lines in PKS1830-211, including thefrequencies are indicated two previously observed, with improved signal-in each panel. The color- to-noise ratio. 28-02-2012ing of corresponding tran- The background source of this system, PKS1830-sitions matches that of ~6 GHz 211, is a high redshift (z = 2.507) blazar, which isFig. 1. The top spectrum 3-1 20E radio loud and time variable and is viewed as twoin each panel is a time- Residuals Kµ = -1 Downloaded from www.sciencemag.org on January 4, 2013 0 spotlike features and an Einstein ring, whichweighted average of the Line-to-Continuum Ratio [%] Combined result from gravitational lensing by the interven-individual spectra, which 08-12-2011 ing face-on spiral galaxy (21, 22). The redshift ofare displayed below the -2.5combined one. For each the main molecular absorptions from the galaxyspectrum, the position of 09-12-2011 is z = 0.88582 (19, 23), corresponding to a look-a fitted Gaussian (depicted 10-12-2011 back time of 7.0 billion years, or half the age ofas light green curves) is the universe (24). More than 30 different molec- 05-04-2012 ular species were detected in the lensing galaxyshown in the graph atthe right. Residuals are of PKS1830-211 (19). Molecular absorption is ~25 GHz 06-04-2012shown at the top of each mostly detected toward one of the two blazar 00 10E, 00 10A+combined spectrum with images (the southwestern), whereas the otherdashed lines indicating Residuals Kµ = -7.4 image (the northeastern) shows weaker andT1s offsets. In the case fewer molecular lines at a slightly different red- 0 Combinedof ~25-GHz observations, shift but stronger neutral hydrogen absorptiontwo proximate methanol [e.g., (19, 25)].transitions were recorded. -1 The CH3OH lines were recorded with theThey are separated by 06-03-2012 100-m single-dish Effelsberg radio telescope,27.494 km/s; the fitted using the 5-, 1.3-, and 1-cm receivers. Preliminarypositions of the weaker detections were performed during the course of 07-03-2012line are corrected to bring 2011, and subsequently systematic observationsthe measurements on a ~32 GHz were performed in a narrow time slot. The datacommon scale. The lines 2-1 10 E were registered onto a local standard of rest ve-are calibrated by the to- -120 -60 0 60 120 6 8 10 12 locity scale, which was centered at z = 0.88582.tal continuum. The blacksquare (upper panel) and Relative Velocity [km s-1] The two blazar images and the Einstein ring arethe black diamond (lower panel) represent the single line observations from Ellingsen et al. (20) and unresolved and PKS1830-211 is effectively treatedMuller et al. (19), respectively. The line positions, originally reported on a heliocentric velocity scale, were as a point source, which is an assumption under-transformed to the LSR scale via VLSR – VHEL = 12.432 km/s. lying the present study. The recorded spectra are shown in Fig. 2. For a single transition, the spectra taken on var- ious days were averaged together, weighting theTable 1. A summary of the relevant parameters and results. Laboratory data: lower and upper energy individual scans by their integration time. Thelevel quantum numbers; laboratory frequencies, n, of the four relevant methanol absorption lines; andtheir uncertainties, fractional uncertainties, and uncertainties in terms of Doppler shift, ∆vD, in km/s. lines were calibrated by the total continuum soCalculations: the sensitivity coefficients, Km. Observations: the measured local standard of rest velocities of that their strength is expressed as line-to-continuumthe lines (relative to z = 0.88582) and the line widths with their 1s uncertainties. Assuming a molecular flux density ratio. The profiles, devoid of under-hydrogen density of 2 × 103 cm−3 and a kinetic temperature of 80 K (12, 29), full width at half maximum lying structure, were fitted as a single Gaussianlinewidths as fitted to observations and a TCMB = 2.728(1 + z) = 5.145 K for the temperature of the (Table 1). The accuracy of the position measure-cosmic microwave background (CMB) radiation at z = 0.88582, the optical depths t yield a total column ments is at the level of 1 to 4% of the line width.density of 2.0 × 1014 cm−2 from a large velocity gradient radiative transfer model (30). The velocities between different transitions are interrelated via Laboratory data Calc. ObservationsLine n ∆vD Position Width V/c = –Km∆m/m ∆n/n Ref. Km t JWKW–JK Ts (GHz) (km/s) (km/s) (km/s)3-1–20 E 12.178597 (4) 3 × 10−7 0.1 (31) –32.8 9.06 T 0.67 16.4 T 1.4 0.0024 where c is the speed of light, and ∆m/m rep-00–10 A+ 48.3724558 (7) 2 × 10−8 0.006 (32) –1 8.40 T 0.10 10.8 T 0.2 0.045 resents the deviation from the current laboratory00–10 E 48.376892 (10) 2 × 10−7 0.06 (33) –1 9.12 T 0.30 14.6 T 0.6 0.016 value of m, defined so that a positive sign indi-2-1–10 E 60.531489 (10) 2 × 10−7 0.06 (33) –7.4 9.83 T 0.43 17.0 T 0.9 0.028 cates a larger m in the high-redshift–absorbing www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 47
REPORTS Fig. 3. The positions of Thus, we obtain a limit on varying m to be ∆m/m = the four observed meth- (–0.1 T 7.6stat T 7.0sys) × 10−8 or, if the statistical anol lines (represented and systematic uncertainties are added in quadra- by V/c with respect to z = ture, a limit of ∆m/m = (0.0 T 1.0) × 10−7. 0.88582) are plotted ver- sus their sensitivity coef- ficients, Km. The bold blue References and Notes horizontal line represents 1. L. Smolin, Physica A 340, 705 (2004). 2. J. D. Barrow, J. Magueijo, Phys. Rev. D Part. Fields the fiducial result of a fit Gravit. Cosmol. 72, 043521 (2005). to the E-type lines, where- 3. J. K. Webb, V. V. Flambaum, C. W. Churchill, M. J. Drinkwater, as the dashed line repre- J. D. Barrow, Phys. Rev. Lett. 82, 884 (1999). sents a fit to all four lines. 4. M. T. Murphy, J. K. Webb, V. V. Flambaum, Mon. Not. R. A positive slope of the fitted Astron. Soc. 345, 609 (2003). line implies that m had a 5. J. K. Webb et al., Phys. Rev. Lett. 107, 191101 (2011). 6. J. A. King et al., Mon. Not. R. Astron. Soc. 422, 3370 smaller value in the early (2012). universe than is measured 7. E. Reinhold et al., Phys. Rev. Lett. 96, 151101 (2006). in the laboratory. The blue- 8. A. Malec et al., Mon. Not. R. Astron. Soc. 403, 1541 shaded surface is a den- (2010). sity plot of simulated data 9. F. van Weerdenburg, M. T. Murphy, A. L. Malec, L. Kaper, W. Ubachs, Phys. Rev. Lett. 106, 180802 (2011). Downloaded from www.sciencemag.org on January 4, 2013 points from the blue fitted 10. J. van Veldhoven et al., Eur. Phys. J. D 31, 337 (2004). line and reflects the con- 11. M. T. Murphy, V. V. Flambaum, S. Muller, C. Henkel, fidence bands of the fit. Science 320, 1611 (2008). Color-coding of the data points is the same as in Figs. 1 and 2. 12. C. Henkel et al., Astron. Astrophys. 500, 725 (2009). 13. N. Kanekar, Astrophys. J. Lett. 728, L12 (2011). 14. P. Jansen, L.-H. Xu, I. Kleiner, W. Ubachs, H. L. Bethlem, Phys. Rev. Lett. 106, 100801 (2011). galaxy (i.e., ∆m = mz – mlab). Therefore, to de- A type methanol should be considered as dif- 15. S. A. Levshakov, M. G. Kozlov, D. Reimers, Astrophys. J. termine the fractional change in m, the peak po- ferent species and thus may undergo spatial seg- 738, 26 (2011). sitions of the four transitions are plotted (in V/c) regation effects. In the combined spectrum, 16. C. C. Lin, J. D. Swalen, Rev. Mod. Phys. 31, 841 (1959). 17. P. Jansen, I. Kleiner, L.-H. Xu, W. Ubachs, H. L. Bethlem, versus Km, and a (dashed) line is fitted to the the 00–10 A+ and 00–10 E transitions, falling in Phys. Rev. A 84, 062505 (2011). data (Fig. 3). Because the A and E levels of close proximity in a single scan of the receiver, 18. F. J. Lovas, J. Phys. Chem. Ref. Data 33, 177 (2004). methanol can be regarded as belonging to two are separated by 0.72 T 0.32 km/s. Moreover, 19. S. Muller et al., Astron. Astrophys. 535, A103 (2011). separate species, the data were analyzed in two the linewidths of the E lines are markedly 20. S. P. Ellingsen, M. A. Voronkov, S. L. Breen, J. E. J. Lovell, Astrophys. J. Lett. 747, L7 (2012). different ways: First, only the three transitions larger than that of the A line (Table 1). Because 21. D. L. Jauncey et al., Nature 352, 132 (1991). from E levels were fitted, then the A transition this is suggestive of a spatial segregation of the E 22. C. Lidman et al., Astrophys. J. 514, L57 (1999). was added to the sample. The analysis of the E and A symmetry methanol molecules, we adopt 23. T. Wiklind, F. Combes, Nature 379, 139 (1996). transitions results in ∆m/m = (–0.1 T 7.6) × 10−8, a fiducial limit on ∆m/m from the fit of only E 24. Adopting a standard L-cosmology with H0 = 73 km s−1 which is consistent with a nonvarying m at the transitions. Mpc−1, Wm = 0.28, WL = 0.72. 25. J. N. Chengalur, A. G. de Bruyn, D. Narasimha, Astron. level of 1.5×10−7 (95% confidence level). The Another source of systematic error is the Astrophys. 343, L79 (1999). reduced chi-squared, cn2, which is a measure of known variability of the lensed object PKS1830- 26. See supplementary materials on Science Online. the quality of the fit, is ~2.0 (26). The fit on all 211. The absorption strength of radio lines was 27. S. Muller, M. Guélin, Astron. Astrophys. 491, 739 (2008). four transitions has a much larger cn2 of 6.4, found to vary strongly, by a factor of >6 in a 28. Each line required a change of receiver, for which reason the time intervals could not be shorter. which might be attributed to segregation is- time span of 3 years, and this was ascribed to 29. K. M. Menten et al., Astron. Astrophys. 492, 725 sues (see below), and it delivers ∆m/m = (11.0 T the intensity changes in the background con- (2008). 6.8) × 10−8. tinuum source (27). This phenomenon might 30. S. Leurini et al., Astron. Astrophys. 422, 573 (2004). The upper limit derived here is statistically cause a varying alignment through parts of the 31. S. M. Breckenridge, S. G. Kukolich, Astrophys. J. 438, 504 (1995). more constraining than that derived in previous absorbing spiral and therewith absorption 32. J. E. M. Heuvel, A. Dymanus, J. Mol. Spectrosc. 45, 282 tests in the radio-domain (11–13, 19, 20). More- through varying Doppler components over time. (1973). over, compared to the methods used in previous Hence this variability may affect the derivation 33. H. S. P. Müller, K. M. Menten, H. Mäder, Astron. Astrophys. studies, the methanol method is more robust of a m-constraint from radio-observations. For 428, 1019 (2004). against systematic effects. In particular, it is much this reason we adopted a measurement strategy Acknowledgments: This work is supported by the Foundation for less sensitive to the assumption that all absorbing to explicitly address the source variability issue. Fundamental Research on Matter program “Broken Mirrors & species reside in the same physical location and Spectra of the anchor lines (the middle panel in Drifting Constants.” H.L.B. acknowledges support from the hence are at the same redshift. Spatial segregation Fig. 2) were recorded in December 2011 and Netherlands Organization for Scientific Research via a VIDI of different absorbers may mimic or hide a var- April 2012, whereas the two strongly shifting grant and by the European Research Council via a Starting Grant. We thank the staff of the Effelsberg radio telescope iation of m. This is the limiting systematic error lines have been observed in-between in Febru- for their hospitality and support. The raw data of the for tests based on the comparison between ary and March 2012 (28). The strong (00–10 A+) radioastronomical observations are available upon request different molecular species, such as the compar- line in the combined spectrum from December from the Max Planck Institute for Radio Astronomy at Bonn ison of ammonia with various rotational lines in 2011 is positioned at 8.32 T 0.10 km/s, and at (email@example.com). HCO+, HCN, CS, and so forth (11–13). The mo- 8.80 T 0.24 km/s in the spectrum from April lecular survey in PKS1830-211 suggests that seg- 2012. The difference between them is 0.48 T Supplementary Materials www.sciencemag.org/cgi/content/full/science.1224898/DC1 regation effects are prominent among different 0.26 km/s, possibly indicative of a small sys- Supplementary Text species (19). For instance, a single methanol line tematic shift due to variability. We have assessed Table S1 was found to be displaced from the average ab- this possible systematic effect as caused by time Reference (34) sorption velocity by more than 3 km/s (19). Our variability in two models (26) and have chosen 18 May 2012; accepted 16 November 2012 test is based exclusively on a single molecular the one producing the largest uncertainty (∆m/m Published online 13 December 2012; species. However, as discussed above, the E and of 7.0 × 10−8) to give a conservative estimate. 10.1126/science.122489848 4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org