Login 
Mass spectrometry in environmental applications
Introduction
In the mid 80s I worked in a research lab which was on the immediate north bank of Melbourne’s Yarra River. The ‘mass spec’ we had there, which was by that time elderly and had been modified several times, had become troublesome to a degree that the Chief of the lab had, he told me himself, been tempted to throw it into the Yarra. Happily that did not happen and, with the aid of an exceptionally skilled instrumentation engineer, my technician and I were able to obtain a few more productive years of service from it.
Mass spectrometry in its various forms provides peaks corresponding to mass number interpretation of which is the basis of analysis. In the work referred to in the previous paragraph, we were concerned with pyrolysis products from coal, and these include carbon monoxide, molecular weight 28. This precluded the use of nitrogen as a background gas, since its molecular weight is also 28, so the response of the mass spectrometer to that would have totally concealed that due to any carbon monoxide formed. Instead we had to use argon, much more expensive than nitrogen, the signal from which was at mass number 40.
Applications of mass spectrometry to environmental matters have been so numerous that lengthy review articles could be written on them and, indeed, have been. This article will comprise selected examples of such use, with emphasis on interpretation of spectra.
Details of particular applications
Information is given in tabular form and comments follow the table. Primary organic aerosols arise from combustion of coal, petroleum products and biomass and once in the atmosphere can undergo further reactions to secondary organic aerosols (SOA), increasing the number and diversity of organic compounds in the atmosphere.
These further reactions involve oxidation, and SOA has the generalised formula CnHmOp. As an example, putting m = 33, H = 56 and p = 3 gives C33H56O3 of mass number of exactly 500. The capability of HR-MS in the analysis of SOA (row one of the table) can be understood in the following way. Resolution of 0.001 mass units is claimed 1 for state of the art HR-MS, and this means that mass numbers are not restricted to whole numbers as they are in the much more basic work outlined in the first paragraph. In the case of C33H56O3 this gives a resolution of: (0.001/500) × 106 parts per million (ppm) = 2 ppm.
Whatever sample of air contains this compound would be likely to contain other ones very close to it in mass number, and unless the peak at 500 is sharp it might merge with the signal for nearby masses and be impossible to identify reliably. This is where the resolving power, usual symbol R, is important. This is defined as: m/?m where m is the mass number of the response and ?m is the signal width at half maximum height. Clearly a large ?m would characterise a spread out and badly defined signal.
Development work in HR-MS has therefore been concerned with obtaining good values for R, and suffice it to say that in the application to SOA, values of R of 20,000 or higher have been realised. To break something down is a means of finding what it is made of, and that was the basis of the work on coal outlined in the first paragraph. Of more interest to readers of this periodical is a similar approach to soil analysis: decomposition under heat (pyrolysis) of soil and identification of products. This features in the entry in the second row of the table, in which pyrolysis products from soil samples heated to 800° C were passed into a mass spec as a molecular beam; that is, as a stream of molecules having a narrow range of velocities and low collision frequency. Whereas in the work on coal referred to argon was, as noted, used as the inert gas in 2, helium was used.
The combined technique of pyrolysis, a molecular beam and mass spectrometry can be summarised py-MBMS. In analysing the contents of the first row of the table we chose mass number 500 as being a reasonable representation, as, indeed, did the authors of 1. In reference 2 the major peaks occur at lower mass numbers than this and the highest values at which peaks were recorded are about 350. For soil not more than 5mm in initial depth from the surface, the most prominent peaks were at about 100 and there is rapid decline in signals up to 150. For soil in the depth increment 15 to 30cm there are many major peaks in the range 110 to 170. The authors of 2 go on to argue that it is ‘new carbon’ recently deposited from the vegetation canopy which has given the characteristic signals in the shallow soil, and that ‘older carbon’, having had more time to decompose, features in the mass spectrum of the deeper soil. This also includes peaks attributed to polyaromatics including very prominent ones at mass numbers 128, 142, 156 and 170.
We can read a little into these in the following way. A mass number of 128 corresponds exactly to naphthalene C10H8. We’d expect that any naphthalene in the pyrolysis products would be accompanied by smaller amounts of substituted naphthalenes; that is, napththalenes with a methyl or ethyl group, leading to mass numbers of 142 and 156 respectively. So this assignment of the peaks appears exact although the authors of 2 do not themselves make it. As for the peak at 170, we note that this is not high enough for the structure to contain three aromatic nuclei as that would require C14 (168) plus some hydrogens. Add to 128 for naphthalene a methyl and an ethyl, respectively 15 and 29, and deduct 2 for the hydrogens which would otherwise have been at the sites occupied by the methyl and ethyl, and the answer is 170, so here again assignment is possible.
Of course, there will be more than one isomer of monomethyl monoethyl naphthalene. Hopefully a reader is becoming aware of how thoughtful interpretation of a mass spectrum can lead to quite precise conclusions. In membrane introduction mass spectrometry (row three of the table) the substance being analysed for is ‘introduced’ into the mass spectrometer by transfer through a membrane. The membrane will commonly be a silicone polymer or possibly of PTFE, and the process of transfer is called pervaporation. In 3 the degradation of chloroform by aqueous sodium hypochlorite was monitored by MIMS, the membrane being composed of a silicone material. Chlorine existing as more than one isotope, there was in the spectrum a peak for CO(35Cl)2 at mass number 98 and a smaller one for the CO37Cl35Cl at mass number 100, and multiple peaks due to the isotopes were observed in other parts of the spectrum.
The reason for monitoring intermediates in the reaction is of course that this provides information on the mechanism, which in turn can lead to control by adjustment of pH and so on. The end products of the reaction are simply carbon dioxide, hydrogen chloride and water. A feature of MIMS is that in the recorded mass spectrum the peak corresponding to the molecular weight of the compound being analysed for will often be accompanied by ones for that weight, plus or minus one. This is because of prior reaction involving hydrogen atoms. As an example we consider acrolein (IUPAC name propenal), structural formula. This is used as a herbicide. It is toxic, and amounts of it in a water sample can be measured my MIMS 4. The molecular weight is 56, yet it is reported 4 that when it is studied by MIMS there are peaks at 55, 56 and 57. That at 57 is attributed to the above structure in protonated form. That at 57 is attributed to subsequent dehydrogenation of such a protonated structure; that is, elimination of a hydrogen molecule from it.
Generalising the discussion and calling the molecular weight of the compound M, the protonated structure will be at mass number (M + 1) and the protonated and dehydrogenated structure at mass number (M – 1). When simple alcohols are analysed in this way the (M – 1) peak tends to be used in characterisation; for ethanol this occurs at mass number 45. In ICP-MS (row four of the table) radio frequency is used to produce a medium (plasma), which is electrically conducting by reason of there being ionised argon and electrons. Residence time in the plasma of a sample for analysis is of the order of a millisecond, and this is sufficient for ionisation of metals present. These are then detected and measured at the instrument with a very good degree of resolution, actually 0.5 units of mass number.
Consequently 5 small amounts of 55Mn can be quantitatively determined in the presence of much larger quantities of 56Fe. ICP-MS can distinguish the two isotopes of silver, 107Ag and 109Ag and also (an application relevant to seawater analysis) can measure 114Cd and 115In in simultaneous occurrence and 208Pb and 209Bi likewise. The exceedingly high temperature of the plasma causes any carbon particles entering it to be vaporised, so these too can be quantitatively determined at the mass spectrometer. ICP-MS is in fact frequently applied to measurement of carbon particles in air. The compound known as dioxin (final row of the table) is actually 2, 3, 7, 8 dibenzo-p-dioxin and its structure is symmetrical. There are many similar compounds grouped with it for the purposes of environmental toxicology and we can correctly talk about dioxins. The molecular formula of 2, 3, 7, 8 dibenzo-p-dioxin itself is C12O2Cl4H4 and the molar mass 322 if the ‘abundance-weighted atomic mass’ of 35.5 is used for chlorine.
The interested reader can very easily access the mass spectrum of 2, 3, 7, 8 dibenzo-p-dioxin by going to reference 6; the mass spectrum will therefore not be reproduced here. We shall, however, interpret it having regard to isotopes and (something which has not featured previously in discussion) peaks due to fragmentation. The strongest peak is that at 322, corresponding to the molar mass. Almost as dominant is the peak at 259, clearly due to fragmentation involving loss of COCl from the original structure. For 35Cl, COCl has mass number (12 + 16 + 35) = 63 which deducted from 322 gives 259. The effects of the isotopes are smaller peaks close to the major ones, and the primary peak at 322 can in fact be interpreted a little more rigorously if such effects are accounted for. The formula C12O235Cl4H4 has molar mass 320 so the one at 322 is, with reference to this formula and using the notation from above, an (M + 2) peak. Not only chlorine but also carbon has more than one isotope.
The structure 13C12O2Cl4H4 would, using 35.5 for chlorine, have molar mass 334. A peak for mass number 334 is in its due place in the mass spectrum being examined 6. Loss of 13CO35Cl from this by fragmentation would give a mass number of (334 – 64) = 270 and this peak too is easily identifiable in the mass spectrum in 6. Hopefully a reader is becoming aware of two things: that the potential for application of mass spectrometry to environmental matters is immense and that very often the mass spectra are ‘transparent’ in that the peaks are so easily correlated with chemical structures. Further application: the mass spectra of two greenhouse gases Given the very high place on the agenda for world affairs occupied by greenhouse gas emissions, a suitable topic with which to close the article is application of mass spectrometry to two greenhouse gases: carbon dioxide and methane. The mass spectrum of carbon dioxide displays a peak at its mass number of 44 as would be expected, and a smaller but quite significant peak at 28 representing fragmentation to carbon monoxide.
There is also a peak at 16 representing the detached oxygen atom. Obviously, if carbon dioxide and carbon monoxide are both being determined the peak at mass number 28 will be due both to the carbon monoxide initially present, and to the fragmentation product of the carbon dioxide. For a particular mass spectrometer the ratio of the peak height at 44 to that 28 due to fragmentation will be the same across operating conditions of interest, enabling the peak at 28 to be corrected for the fragmentation effect.
Methane is of course a more powerful greenhouse gas than carbon dioxide, which is why there is currently concern about release of methane from ‘natural gas hydrates’ in polar regions. By far the dominant peak in the mass spectrum of methane is that at 16, corresponding to the molecular mass. There is also a peak at 15 due to fragmentation to CH3; there is no fragmentation beyond this to CH2.
Concluding remarks
Applications of mass spectrometry become more and more advanced through ongoing development, and this is of course true of any major analytical technique. An appealing feature of mass spectrometry is that easily understood mass numbers feature so centrally. Interpretation according to fragmentation patterns and, in some cases, the existence of more than one isotope, has been demonstrated in this article which, it is hoped, will equip a reader with sufficient know-how and confidence to understand a mass spectrum.
References
1. http://pubs.rsc.org/en/content/articlepdf/2011/cp/c0cp02032j
2. http://www.anl.gov/PCS/acsfuel/preprint%20archive/Files/47_1_Orlando_03-02_0029.pdf
3. http://www.scielo.br/scielo.php?pid=s0103-50532005000200021&script=sci_arttext
4. http://www.sciencedirect.com/science/article/pii/003991409280209V
5. Montaser A (Ed) Inductively Coupled Plasma Mass Spectrometry Wiley-VCH (1998)
6. Reiner EJ ‘The analysis of dioxins and related compounds’ Mass Spectrometry Reviews 29 526-559 (2010)
Author
Clifford Jones, DSc FIChemE FRSC FInstP FIMechE, University of Aberdeen Clifford Jones is a Reader in the Department of Engineering at the University of Aberdeen. He has held academic posts in Australia and the UK and is the author of 12 university-level textbooks, with a 13th in press. He has lectured on his works in countries including the US, Thailand, South Africa, Sweden, Spain and India and also has significant broadcasting experience.
Published: 01st Mar 2012 in AWE International
