The discovery of chemically produced mass independent isotope effects: The physical chemistry basis and applications to the early solar system, planetary atmospheres, and the origin of life

Abstract

I wish to begin by thanking the people who nominated me for the Leonard medal and the Leonard Medal Committee of the Meteoritical Society for granting me this award. Thanks as well to Francois Robert for doing a fine job on the citation and his kind words. I have been a member of the Meteoritical Society and attended the meetings since I was a postdoctoral fellow at Chicago. The recognition by the society means a great deal to me and it is an honor to receive the Leonard Medal. I know most of the recipients over many decades and I am being included with a prestigious group. My research career in Meteoritics and Planetary Science began with my going to Chicago to work with Bob Clayton on nitrogen isotopes in the solar wind as detected by extraction from lunar samples. It was a lucky break for me that Bob had an opening for a post doc in the window of time when I was graduating and looking. Chicago was, and is, a center of research activity in meteorites and planetary science. My interest in the field dates to my last course at the University of Miami when I took a course titled “Modern Geology” taught by Cesare Emiliani who was a student of Urey's. I had taken all the available physics and chemistry courses and wanted another science course. My interest in geology dated back to my childhood, when I belonged to the Gem and Mineral club at Washington University. This was a first-class organization wherein I learned a lot and attended monthly lectures with my dad and the whole family went on the many field trips they organized. Emiliani's course turned out to be the best course I took and he was an amazing teacher and mentor. It was in this broad ranging course that I learned of the application of both stable and radioactive isotopes to interpret processes in nature, ranging from the age of the Earth to paleoclimates, a field he helped found. It greatly impressed me that one could apply the fundamental physics and chemistry I had learned to quantitatively interpret differing geochemical records. I was struck by Urey's 1947 paper on the thermodynamics of isotopic substances and how one could use the information for. Emiliani brought me to his laboratory and showed me his mass spectrometer built at the University of Chicago before he came to Miami. A second significant impact on my scientific direction resulted from a course I took at Florida State from Gary Brass. He discussed the famous paper of Clayton et al. (1973) on the discovery of oxygen mass independent isotopic anomalies in meteoritic inclusions. I was very impressed with the fact that one could distinguish between nuclear processes and chemical by measurement of the multi-isotopes of oxygen and potentially identify individual grains that may have predated the solar system. As I neared completion of my doctoral work, which was finished at Brookhaven National Laboratory, I wrote to Bob Clayton about the possibility of postdoctoral work with him. I told him that I had experience in measuring oxygen and sulfur isotopes, as well as accelerator use for my PhD dissertation. He had an opening and I was able to join the Chicago group, which I had heard about from Emiliani when he discussed his own classmates’ research (Wasserburg, Craig, Epstein, and Miller) and his work with Urey. My timing for coming to Chicago for meteoritic research could not have been better. Bob Clayton, Larry Grossman, Ed Anders, and Frank Richter were all engaged in meteoritic research; all of whom became Leonard medalists. It provided the opportunity to sit in on graduate courses by S. Chandrasekhar and Dave Schramm as well as geophysical, physical chemical, and astrophysical seminars by the world's best. The timing was also perfect as a large group of others of my own vintage joined, including Glenn MacPherson, Andy Davis, Typhoon Lee, Alan Mathews, Miriam Bar-Mathews, Ian Hutcheon, Tim Vander Wood, Richard Becker, and numerous others. They were a great group to interact with and remain so except, sadly for Ian Hutcheon. As was true for many of us, I was also fortunate to have the opportunity to work with Tosh Mayeda, originally hired by Sam Epstein to run Harold Urey's laboratory and then with Bob when Urey came to UCSD. She had taught laboratory research techniques to many before me including many legendary figures such as Harmon Craig, Stanley Miller, Jerry Wasserburg, and my undergraduate mentor Cesare Emiliani. Everyone who has passed through the doors of the Clayton laboratory know how important Tosh was for the laboratory, and in helping everyone with whatever science they were doing including myself. The overall environment was a rich one, especially for meteoritics, and it was a fantastic time to be able to have such a group to interact then and over the course of my career. In my third year at Chicago, the University of California San Diego announced a position for cosmochemistry in the Department of Chemistry to replace Hans Suess, one of the founders of cosmochemistry and well known for many discoveries. His grandfather, Edwin Suess was a professor at the University of Vienna and a geologist as was his father. Han's work was on the origin of the elements and meteorites, radioactivity in K, 14C development, climate studies and models for future global warming, photochemistry of the upper atmosphere, and geophysics, and also a Leonard medalist. His work on elemental origins and the appearance of the “magic numbers” in 1949 and Maria Mayer's breakthrough was of fundamental importance. Maria Mayer was later one of UCSD's first physicists and won the Nobel Prize in physics while at UCSD for the work she did at Chicago. I was invited to UCSD for an interview by the chair of the committee, Jim Arnold, also a Leonard medalist. I was again fortunate and UCSD hired me as a beginning assistant professor. My new colleague and later good friend Stanley Miller mentioned to me that I was to replace Hans Suess. Furthermore, I would inherit the laboratory and mass spectrometer of Urey, which he brought to UCSD when he was hired as one of the first professors here. This was Stanley's typically quiet way of saying that the expectation level was high. The group of colleagues at UCSD was not unlike Chicago. Hans Suess was still active and in an office next to me and I was able to have near daily conversations with him for many years which I very much appreciated. His knowledge of a broad range of science and the history of science back several generations was quite unique. Hans Suess, Jim Arnold, Kurt Marti, and Guenther Lugmair who were along the same corridor as my office and laboratory and Stanley Miller was nearby. In my own age group, Kuni Nishiizumi, Candace Kohl, and Peter Englert were there as well. A steady stream of visitors also coming to work with faculty colleagues, especially during the winter when there was a steady stream of snow evaders. Hans and Ruth Suess had many international visitors and my wife Nasrin and I were invited on many occasions to come to their home to meet them. With Scripps Institution of Oceanography a short distance away (two blocks from my apartment), there was a steady stream of visitors to discuss science with. With Harmon Craig, Davendra Lal, Miriam Kastner, V Ramanathan, and Dave Keeling on the faculty there, many people were coming through to visit and collaborate with. As with all new faculty, the candidate new professor presents a research prospectus as to what they would do. In my case, I discussed two directions; one was to continue and expand the work on nitrogen isotopes on the Moon and meteorites and the second was to test the assumption that all isotope effects must be mass dependent and determine if a chemical mechanism was possible. After reading the Clayton paper many times after Brass discussed it, I had always thought it was a major assumption and worth further consideration and experimental testing of it. Since oxygen is the major element in stony planets and asteroids, the effect in CAI is extraordinarily large. Nearly all stony meteorites are mass independent at the bulk level, meaning the source of the effect must be a major event in solar system formation. The testing of the assumption was not trivial as doing reactions that involve relevant gas phase reactions are not easily done and are experimentally difficult to cleanly do at the isotope level. Other than early work by Gustaf Arrhenius at Scripps, who tried some elegant experiments to look for effects, there was nothing to base experiments on. My choice of ozone formation was based upon several factors. The quantitative separation of products from reactants and avoidance of secondary reactions made it an attractive choice experimentally. A second factor was derived from a comment made to me at Florida State University where I did my PhD I was working on the 20 MeV accelerator there, and my office housed in the Nuclear Research Building (NRB) of the physics department. A famous physicist named P.A.M. Dirac retired to FSU and was in the physics building connected to NRB. He once made a comment to me about the importance of symmetry in governing various processes in physics and the adherence or violation of it. From that vantage point, ozone is unique in that it is one of few gas phase reactions where the product molecule is composed of one element. It has the requisite criteria of possessing three or more isotopes. This minimum number of isotopes is required to determine if a process is mass independent. It is a very special molecule from a symmetry vantage point of view. Finally, the ozone experiments were also experimentally possible within my financial domain. The only snag was that I needed to modify Urey's old mass spectrometer to do all three oxygen isotopes. With council from Guenther Lugmair, I remade the collector system to measure O2 for all three oxygen isotopes. Figure 1 shows the results of the first experiments on ozone formation that my first student, John Heidenreich, and I did on ozone (Thiemens and Heidenreich 1983). Figure 1 shows that product ozone is equally enriched in 17O and 18O, i.e., it forms a slope of 1 in three-isotope space and is essentially the same as that observed in CAI, with a plot of the CAI data of Clayton et al. (1973) shown as well. The identicality of the meteorite and laboratory isotopic composition experiments is striking and begs the question of the relation of the two. When we were considering the observed phenomena in Fig. 1, though a very simple relation, there was zero theory present in chemical physics that could account for the observations and it took a lot of investigation to determine the causal mechanism. In Thiemens and Heidenreich, we suggested isotopic self-shielding, pointing to the “variations observed in the isotope ratios of CO isotopes in the surface layers of molecular clouds due to isotope selective photo destruction of CO.” A slope one might arise because the fractionation is based upon natural abundance differences and not a mass effect. Ralph Cicerone had also suggested self-shielding (Cicerone and McCrumb 1980) in the Earth's atmosphere in molecular oxygen because of doubling of the rotational states of 17O18O and 16O18O compared to 16O16O, a symmetry effect and in fact how oxygen isotopes were discovered (Giauque and Johnston 1929a, 1929b). I had in fact stopped at the National Center for Atmospheric Chemistry, en route to San Diego from Chicago to talk about his work. Ralph became a good friend and was later the chancellor at UC Irvine and then president of the National Academy of Sciences. The line doubling is an important facet and was an early proof of quantum mechanics. In a graduate course on quantum theory, I was impressed by this discovery of oxygen isotopes and the underlying physics. They reported that in the atmospheric absorption spectra of oxygen in the Earth's atmosphere, they found that weak lines in the spectra could not be accounted for by the strong A and weak A′ bands and that it must be due to the presence of an isotope of oxygen. The appearance of alternate rotational lines for the 17O16O, 18O16O versus 16O16O species (twice as many states) is a quantum symmetry effect. In our manuscript, we suggested that symmetry factors might also play a role, but at the time, there was no study or theory for what part of the reaction processes (or dissociative process) they occur at. In a later paper (Heidenreich and Thiemens 1986), we developed a basic physical chemical theory where we recognized as key the doubling of the states for the stabilization of the very short-lived transition state, *O3 formed immediately following the O + O2 reaction. It was because of the doubling of states for 17O and 18O in the asymmetric species providing more access to the ground state created the slope 1 and mass independence. During the short time scale of this transition state, it may either stabilize to O3 or re-dissociate to O + O2. The stabilization process is a minor pathway and most transient species re-dissociate. The probability of stabilization relies in part upon the transient lifetime, with longer lifetimes resulting in greater stabilization probability. The lifetime is partially moderated by the number of available states and we concluded that the state doubling leads to a great probability of stabilization for the asymmetric species compared to the symmetry, consequently producing the observed slope 1 effect with heavy isotope enrichment in ozone. An outcome of a symmetry effect is that it is general and not restricted to ozone. Furthermore, it accounts for the observation that oxygen is the only element that exhibits anomalous isotopic behavior at the bulk level. Oxygen, due to its position on the periodic chart, coordinates elements and geochemical building blocks (SiO4, Fe2O3, and Al2O3) is subject to symmetry mediation. Carbon and hydrogen may also be subject to symmetry selection, but only possesses two stable isotopes and consequently mass independence cannot be tested. Sulfur is a possibility and symmetry effects may be of relevance in the atmosphere and Archean. The precise quantum mechanical details of the source of the anomaly have not been developed to include all parameters, though all models maintain a symmetry dependence. Important advancement in the theoretical basis of the effect is by Nobel Laureate Rudy Marcus (Hathorn and Marcus 1999, 2000; Gao and Marcus 2001, 2002; Marcus 2004, 2013). These works consider a quantum dynamical approach with intramolecular dynamics and the symmetry of the molecule effects upon transient lifetimes and their role in energy exchange as well as entry to exit channels. In the theory, the isotope enrichments, rate constants for the different isotopomers, and pressure dependence are included. Mauersberger et al. (2005) and Thiemens (2006) provided a review of the measurements and the dependency of the ozone formation process on different parameters. A different theoretical approach has been developed by Babikov and colleagues (Ivanov and Babikov 2009; Telukhin et al. 2013) and utilizes a treatment of the ozone formation process within a framework of mixed quantum/classical dynamics. Scattering resonances become important with this approach, and there is a dependence of lifetimes on rotational excitation and symmetry. The models also account for the rate constants of the isotopically substituted species and pressure dependencies. At present, predicting the magnitude of fractionation is not possible by any model though recent work on sulfur reactions has utilized an approximation for what the pure effect may look like isotopically (Babikov 2017). As regards applications to meteoritics, the models all support the role of symmetry in chemical reactions and Marcus (2004) has made a model that provides a chemical mechanistic source of the oxygen anomalies on an evolving grain surface. In the original paper (Thiemens and Heidenreich 1983), we suggested self-shielding in the early nebula with a UV active Sun to account for the oxygen isotopic composition in meteorites, a theory later resurrected by Clayton (2002), Lyons and Young (2005), Yurimoto and Kuramoto (2004), Sakamoto et al. (2007), Young (2007), and Lyons et al. (2009). The new models include an environment that traps the photoproducts prior to isotopic exchange. Navon and Wasserburg (1985) demonstrated that exchange would remove the isotope signal and cold trapping as water ice is a possible means to protect it. A common feature of the papers is that the cross sections must be precisely determined for self-shielding modeling. The shielding process involves filtration of light with passage through a gas, with selective removal of isotopic lines based upon abundance of the isotopically substituted species the light passes through. A consequence is that the rapid attenuation of the most abundant 12C16O and preferential dissociation of 12C17O and 12C18O. Since the dissociation is a function of abundance rather than mass, the slope is 1, rather than ½, a point we made in the 1983 paper. An issue for self-shielding had been that following dissociation, there would be an immediate loss due to exchange with CO of the anomalous atomic oxygen (Navon and Wasserburg 1985). The problem is circumvented using UV light of the local interstellar medium and irradiating the disk edge where low temperatures prevail allowing the product oxygen to be trapped as water (Lyons and Young 2005). The water must be sequestered, transported, and ultimately converted to silicates in a very precisely and highly balanced mixing to create the observed meteoritic oxygen isotopic species. This assumes no photodissociation of the water itself and no isotope effect during dissociation. The self-shielding mechanism requires that dissociation occur by the UV absorption and the spectral lines must be separate. To create the slope 1, it is absolutely required that the cross section for the C17O and C18O are the same, especially given that (1) the cross section enters in the exponent of the absorption (e−σcl, where σ is the cross section and cl is the column density) and (2) it involves the minor abundance species. Given that there were no experimental tests of self-shielding and the cross sections not measured at the precision needed for models are limited, we experimentally tested the models. Even a minuscule difference in the cross sections of the lines for the minor isotopes would result in a slope of non-unity since the exponential term is so sensitive. In addition, and most importantly, as was the case in the supernovae theory, a very fundamental assumption made which is that following the optical shielding, there is no other isotope effect and the dissociation process does not alter the effect of shielding. This was an untested hypothesis and Subrata Chakraborty and I set out to test self-shielding of CO. The testing turns out to be difficult as the UV range operative in the nebula is sufficiently short that there are no photocell windows transparent at these wavelengths. There are also no laboratory light sources and we needed to develop experiments from scratch at a synchrotron. Subrata in collaboration with Musa Ahmed at the Advanced Light Source (ALS) Lawrence Berkeley laboratory developed an experimental device that used a windowless system with CO gas flow and trapping to remove the product atomic oxygen for isotopic measurement at UCSD. With the complexity of the system and collection, interfacing to the synchrotron, and isotopic analysis of sub-micromolar amounts of oxygen, these were the most difficult experiments we have done to date, including developing a rocket-borne payload to sample stratospheric and mesospheric air at White Sands Missile Range for return laboratory isotopic analysis (Thiemens et al. 1995). The experiments tested self-shielding by executing photolysis at multiple wavelengths to access various electronic states to determine the role of photodynamics and dissociation selectivity and the associated isotopic fractionation. This had not been tested for any molecule at the multi-isotope level. As discussed, self-shielding requires that the isotopically substituted species absorb light at different wavelengths. As light passes through the gas, C16O, C17O, and C18O absorb at wavelengths different from one another due to differences in zero point vibrational energy. Consequently, the most abundant isotope absorbs light preferentially as the absorption, given in the equation in the foregoing paragraph depends upon cross section, path length, and concentration. The cross sections and path length are assumed the same in a passage of light through a gas. The differences in isotopic absorption rate arise from the natural abundances of the various isotopes. C16O (natural abundance 99.76%) absorbs shielding light over a path length greater than C18O (0.20%) and C17O (0.04%). With increasing distance, C17O and C18O molecules preferentially dissociate (Thiemens and Heidenreich 1983). There exists a slice in distance, IF the cross sections are exactly equal; they produce a slope 1 in the three-isotope plot in Fig. 1. This occurs in a select path length and varies in nonunity values before and after this zone. The direct test of self-shielding was done at wavelengths of solar nebular relevance. These wavelengths correspond to 105.17 and 107.61 nm. The UV regions in the nebula where self-shielding occur are restricted due to the interference of the more abundant H2 gas, which absorbs most available UV light. It is predominantly at 105.17 nm that self-shielding might occur. It is in a hydrogen window and has the requisite separation of the isotopic lines of CO required for shielding. At 107.61 nm, dissociation occurs but there are no isotopically distinct lines. Without isotopic separation, self-shielding cannot occur since there is no differential path length difference in isotope dissociation rates. In Chakraborty et al. (2008), we reported that there is a massive effect (up to 2430‰ in δ18O) at 107 nm. At 105.17, we did not observe the requisite slope 1 in the isotopes that would prove shielding. The results at 105 and 107 nm were identical in isotope slope inconsistent with self-shielding. A slope 1 must be produced at 105 and ½ at 107. The wavelength dependency shows that the dissociation process is via a mechanism not attributable to self-shielding. The paper published in Science drew a lot of comment. Since the experiments did not agree with models, the experiments must be wrong; consequently, we pushed the measurements further, testing other aspects of dissociation, and provided deeper evidence on the CO dissociation process. We rebuilt the synchrotron photo apparatus and added an ability to control the temperature and test its effect. Duplicate experiments resolve any issue of artifacts and we measured across wavelength to resolve the fractionation versus wavelength. As a further test, the pressure (column density), which is the main factor in shielding, was varied by nearly an order of magnitude. In Chakraborty et al. (2012), we reported that there is no effect in pressure (columnar shielding) on the oxygen isotope slope where 1 is expected. The temperature variation of the photolysis was altered in the three-isotope fractionation slope from 0.75 to 1.1, inconsistent with a shielding effect. More importantly, the experimental observations are consistent with the effect arising after the absorption of light and during the process of dissociation. The consequence is that even if there is a filtration of light by self-shielding, there is a isotopic selective process that overwrites the shielding (Chakraborty et al. 2018). There is a crossing between electronic states prior to dissociating and isotopically highly selective. The observed temperature effect significantly alters the dissociation process and its selectivity. Chakraborty et al. (2018) for the first time measured the carbon isotopes in the process. The cross sections for 13CO are well known and the self-shielding for the experiment may be precisely calculated since all parameters are precisely known, including the beam profile and precise light flux across the beam profile. The experimental results for carbon isotopes are inconsistent with self-shielding and do not follow what is expected for the carbon in the CO experiments. The experiments on CO have shown that the use of isotopes in studying detailed quantum chemical processes is a new tool and gives insight that is unobtainable from other physical chemical methods. Since our first experiments, the basic theme of our approach is to tackle a fundamental physical chemistry phenomena, resolve its mechanism, and then apply it to different aspects of natural processes. The process of CO dissociation at the isotope level is a good case in point. In our attempt to push forward in understanding the very basic details of photodissociation at the isotope level, it became clear that there are many basic physical chemical phenomena associated with dissociation that are simply not known. The same is true for the reaction theory associated with the isotope effect for ozone formation and symmetry considerations. What we have reported is that there are selection controls on reaction processes observed at the isotope level that are fundamental and only apparent with isotopes. Our published experiments are explainable on the basis that the effect is not a direct consequence of either the primary photon absorption process or a simple relation with the bond strength. Rather the selection process occurs during the crossing between excited states prior to dissociation and in the dissociation dynamics. This is a highly complex area of physical chemistry being intensively studied and a state of the art topic in chemical physics. I have been fortunate to be able to discuss our results with many in the field since part of my PhD work was in photochemistry. I also chanced upon good fortune that greatly advanced our development of photodynamic theory. During a visit to Hebrew University, I gave a seminar and it was a nice opportunity to revisit with Alan and Mira Mathews, Yehoshua Kolodny, Boaz Luz, and Oded Navon. I had asked if they might introduce me to Raphy Levine, regarded as one of the very best theoretic quantum dynamics and photodissociation theoreticians in the world. I am still appreciative that they did; he came to my talk to learn a bit about isotope applications and I subsequently had a chance to talk with him. I asked Raphy about the photochemistry of CO and explained why I was interested in it and the results of our experiments. His comment was that from a theoretical perspective, one cannot fully track the photodissociation process at the isotope level in CO because the surfaces are not sufficiently well known and not all the isotopic cross sections are measured. Furthermore, due to perturbations among the excited states, the cross sections may not be calculated and must be precisely measured. His suggestion was that since N2 is isoelectronic with CO and all the cross sections of the isotopic species have been carefully measured, it was a good molecule to study as an analog. One can answer the basic questions for isotope effects in CO dissociation using nitrogen and determine what the relevant isotope factors are. I worked with Raphy and his colleagues on addressing the question as to what the isotope effect is when a simple molecule absorbs a photon and dissociates and what the concomitant isotopic effect is. What we found is that the dissociation process is extraordinarily selective and it is in the coupling of diabatic electronic states of different bonding character that occurs after the excitation of these states (Muskatel et al. 2011). It is well known that this coupling is characteristic of energy regimes where two excited states are nearly crossing, such as for N2 and CO. The effect is simultaneously large and highly wavelength-dependent and these massive effects may not arise from self-shielding. To define the isotopic photodynamics at higher resolution and the electronic state, crossing the dynamics feature that is the source of the fractionation, it was necessary to find a way to develop a highly specific experimental approach. I therefore spent time in Jerusalem discussing this point with Raphy. After a few days of discussion, Raphy arrived at a prediction and a specific test. Comparison of the calculated and measured cross sections show for nitrogen there are wavelengths where there is significant disagreement. The source of the disagreement is a consequence of the excited electronic state mixing at such wavelengths and state of the art models cannot capture this interplay. If the isotopic selection results from state to state crossing at these perturbed wavelengths, there should be a maximal isotope effect. The wavelength where disagreement is most noticeable and maxima of states interaction are at 90 nm. At this wavelength, the isotope effect be anomalous and provide a direct test. It is observed that across UV wavelengths, the basic isotope effect associated with simple dissociation of nitrogen is greater than a δ15N of 2000‰ (Chakraborty et al. 2014). The key test at 90 nm shows a massive enrichment that rises to greater than 10,000‰ confirming that the selectivity arises from the state mixing. The results are displayed in Fig. 2. Another advantage of nitrogen in the experiment is that all of the isotopic lines are measured, the exact column density is measured, and the flux of light as a function of wavelength is measured. Calculations for the isotope effect associated with self-shielding are done at high precision and accuracy. In Chakraborty et al. (2014), a comparison between measured and calculated differs by thousands per mil. The massive spike observed at 90 nm in the experiments is not present in shielding models.