Atmospheric Pressure Ionisation of Small Molecules

Skip to Navigation

Education Article

  • Published: Jan 15, 2014
  • Author: Tony Mallet
  • Channels: Base Peak
thumbnail image: Atmospheric Pressure Ionisation of Small Molecules

Professor Tony Mallet. School of Science, University of Greenwich, Chatham Maritime, Kent ME4 4TB, UK.


Since the early 1980s many different and ingenious attempts have been made to achieve reliable and sensitive ionisation at atmospheric pressure in order to provide mass spectrometric data from molecules which were too involatile and thermally unstable to be submitted to probe or GC analysis. Fast Atom Bombardment was the first really satisfactory approach but was not ideally suited to linking to an HPLC eluant. Direct inlet devices could only accept very low flows and blocked at the slightest provocation while moving belt interfaces were also very temperamental in their behaviour. Thermospray ionisation addressed many of these drawbacks but the real resolution of the problem arrived with electrospray (ESI) and atmospheric pressure chemical ionisation (ApCI). This article will address the application of ESI, ApCI and, to a lesser extent, the recent development of photoionisation (APPI), and direct desorption spray ionisation methods, as they are applied to small molecule analysis and is an attempt to show how many different classes of molecule can be examined by these techniques.

Principles of ESI ionisation (ESI)

Electrospray Ionisation (ESI) has now been in routine use for nearly two decades since the pioneering work of Dole in 1968 and 1971 and the application to mass spectrometry in 1984 by the 2003 Nobel Prize winner, John Fenn. A recent trawl in the on-line Medline database gives over 12,000 references to the search term Electrospray mass spectrometry and this accurately reflects its popularity for analysis of molecules of biological and biomedical interest. Many of these references will be to problems in proteomics and to the characterisation of large polymeric molecules but many will also refer to the ionisation of smaller molecules.

In spite of its well established role and use in mass spectrometry there are still areas of its behaviour which are ill understood and the ground rules which govern whether a given compound will or will not be ionised by the process can be obscure.

The Electrospray source is simple in its design and implementation but the underlying physical chemistry can be complex and is still not perfectly understood. What is required is a method of producing a fine electrically charged aerosol from a dilute solution of the analyte. The solvent should be volatile and often needs to be buffered to a defined pH, using volatile buffers.

Most ESI sources may have a form of concentric capillary tubes down which are fed a solution of the analyte, and flows of nitrogen gas to aid the formation of fine sprays. The capillary carrying the solution is held at a high voltage relative to the inlet into the mass spectrometer. When very low flows of small samples is employed ('nano-electrospray') a quartz or similar capillary holding ca. 1 µl solution, with a tip internal diameter of <10 µm is used and no extra nitrogen gas flows are needed.

Figure 1

Figure 1. A commercial ESI source. Note that here the spray is formed orthogonally to the mass spectrometer
inlet which prevents much of the uncharged species present entering the mass spectrometer.

The influence of a high voltage applied to a capillary filled with a solution, is sufficient in itself to create an electrospray of charged droplets. Solvent evaporation from the charged droplets leads to increased coulombic repulsion and eventual fission, when the Rayleigh limit is reached, into smaller droplets.

Figure 2

Figure 2. The picture above the budding of smaller droplets as the charge density on the drop reaches the Rayleigh limit.

It has been calculated that each successive smaller droplet carries 2% of the parent mass but 15% of the original charge. Eventually a single charged ion related to the original analyte molecule remains. Design of the ESI source needs to include both extra nitrogen gas flows and a some form of heat to assist the evaporation process if flows above a few microlitres a minute are to be accommodated. In practice the lower the flow rate can be held the more efficient and sensitive the ionisation process will be.

Eventually a family of ions virtually free of solvent, carrying either a positive or a negative charge, will result. These ions have to be introduced into the high vacuum of the mass spectrometer analyser. This is achieved by directing them through one or more fine apertures or through a capillary with one or more stages of vacuum pumping. To prevent neutral species passing into the instrument an orthogonal arrangement in which the ESI spray is produced at right angles to the initial mass spectrometer aperture is frequently employed as was shown in Figure 1 above.

The design of the atmospheric pressure interface to the mass spectrometer analyser is not discussed here; suffice it say that considerable care needs to be taken to permit the maximum yield of ions to pass into the instrument and to overcome the powerful cooling effect of passing the spray from atmospheric pressure into a high vacuum, through a small aperture, a process which leads to the formation of solvent cluster species of extremely high mass. In the course of the ionisation process the environment of the molecules in the droplets will be altering drastically over the 100 microseconds that it takes to produce a gas phase species from one in a classical solvent phase. In particular the pH and ionic strengths will change significantly and this can affect the nature of the final ionic species formed. If a mixture of substances is present in the spray those which will tend to accumulate at the surface of the droplets, i.e. the more lipophilic, will be ionised preferentially and may suppress partially or totally the ionisation those substances which remain further inside the evaporating droplets. The consequence of this is that trouble should be taken to ensure that substances are admitted to the source in as pure a fashion as possible and that experiments need to be undertaken, specially when attempting quantitative analysis by, say stable isotope dilution assays, to determine the extent of analyte signal suppression during an HPLC run. Common solvents contain many impurities and a helpful list of the ESI and ApCI responses to these has been published.

One aspect of an electrospray source, which is not necessarily immediately obvious is that it forms an electrochemical cell.

Figure 3

Figure 3. An ESI Source as an electrochemical cell.

In positive ion mode the flow of positive charge to the mass spectrometer is balanced by an opposite flow of electrons from the needle through the high voltage supply. Oxidation takes place at the capillary needle. In negative ion mode the opposite flows occur. If a metal capillary is used with a sufficiently low oxidation potential then the following reaction can occur:

Equation 1

Any hydroxyl ions present can also be oxidised as in the reaction below:

Equation 2 

Observations of the adventitious oxidation of analytes have been reported but several authors have described experiments where the phenomenon has been specifically tailored to enhance the analysis of otherwise intractable materials, one such application is described below. There are a number of reviews of electrochemical applications of ESI Mass Spectrometry.

Atmospheric pressure Chemical Ionisation (ApCI)

In an ApCI source a very similar arrangement of flowing volatile solvent containing the analyte is used but no voltage is applied to the solvent capillary. Instead a fine needle is introduced in the spray and a high voltage is applied to it; this creates a plasma in the source with high energy ions such as N4+° and O2-° from the air, water and solvent. When typical reversed phase HPLC solvents are employed the principle secondary chemical reagents are H3O+ and OH- and these react with the analyte to create a true chemical ionisation. If non aqueous solvents are used chemical ionisation will also take but of course with different reagents. Again a heat source is also included in the design and high temperatures are applied to the end of the capillary. Several reports exist showing that for some analytes a 'Thermospray' type of ionisation takes place and the plasma voltage can be left switched off.

Whereas ESI works best with low solvent flows, ApCI sources optimise at flows around 500 µl/min - 1.0 ml/min. This makes the technique more accommodating to 'classical' analytical HPLC separation methods. A further advantage of the process is that it is reported to be less sensitive to the presence of relatively high concentrations of common buffer compounds.

Figure 4

Figure 4. A schematic for an ApCI source.

A specific application of the ApCI source arises from the observation that it contains a large population of thermal electrons. These are well known as the ionising principle in electron capture GC and GCMS. For appropriate chemical structures with high electron capture cross sections very efficient ionisation can take place. Two examples of applications of this are detailed below.

Atmospheric pressure Photoionisation

The pioneering work of Andries Bruins and colleagues has led to the production of an atmospheric pressure source in which the ionisation is effected by the use of photons from a UV source. These first interact with a 'dopant' which is a UV energy absorbing substance forming a chemical ionising species which in turn reacts with the analyte. The principal virtue claimed for this source is that it can ionise those very non-polar materials which are not detected in ESI or ApCI sources. The availability of commercial APPI sources is still new and while encouraging results have been reported, the verdict on the applicability of the technique still needs more results. A recent article has compared the performance of ESI, ApCI and APPI on a set of aromatic compounds.

Figure 5

Figure 5. Diagram of a Photoionisation source.

Two ionisation processes are possible with the APPI source. In one the dopant e.g. toluene forms a benzyl radical and if this has a lower proton affinity than that of the analyte the latter will form a protonated cation. If, however, the electron affinity of the toluene radical cation is higher than the electron affinity of the analyte, the latter will be seen as a radical cation.

Direct Spray Desorption Ion Sources

Recently a number of ionisation methods which rely on the combined desorption and ionisation directly from surfaces have been described. These have the advantage of producing mass spectra directly from surfaces held in the open atmosphere almost instantaneously. The first sources were a commercial implementation called Direct Analysis in Real Time (DART, Jeol) and Direct Electrospray Ionisation from the laboratory of Graham Cooks. In the latter an Electrospray source is employed to direct a jet of solvent vapour droplets from a charged capillary onto a surface on or in which the analyte is contained. Typically this can be a tlc plate, a filter paper a PTFE membrane onto which a solution has been applied or even a commercial drug tablet. In DESI the spectra are similar to those obtained by 'classical ESI' with protonated (or in negative ion de-protonated) ions predominating but with a wider range of polarity of compounds which ionise efficiently. The DART system relies on the interaction of an ionised gas with atmospheric vapour to lead to similar spectra. A number of variants of the technique have been described, Direct Atmospheric Pressure Atmospheric Chemical Ionisation (DAPCI and ASAP) where the voltage on the plasma needle can be turned on or off, Jet Desorption Ionisation (JeDI) where a high velocity electrically charged water jet acts as the desorption-ionisation agent and Electrospray Assisted Laser Desorption Ionisation (ELDI) where desorption takes place by the action of a laser beam and a stream of charged ESI vapour is used to ionise the product.

Illustration 1

As many of the spectra arising from these direct desorption methods will contain ions from several components the additional use of tandem mass spectrometry is almost essential if confident assignments are to be made.

Examples of the analysis of different compound classes

It is a truism that, in order for a molecule to ionise in an ESI source it needs to be sufficiently basic for it to be protonated (or cationised) in a positive ESI source or sufficiently acidic to lose a proton in a negatively charged ESI source. The problem in predicting whether any given substance will or will not ionise lies in the lack of thermodynamic information about the environment in the gaseous phase leading to formation of the ionised species. Almost any organic compound containing a covalent nitrogen and most carboxylic acids, as well as those substances which exist already ionised, will be detected with good sensitivity by this source. It is molecules containing other heterogeneous atoms such as oxygen and sulphur etc which are harder to predict. When ESI fails then ApCI can be tried and in both cases both polarities need to be attempted. Compounds which are already ionised are obviously the easiest to analyse and these include those with sulphate and phosphate groups, those containing quaternary ammonium moieties and many acids which easily form anions. Most nitrogen containing compounds with primary or secondary or tertiary amines as well as amides are easily protonated. In this section a number of different compound classes are discussed and methods which have been employed for their ionisation are illustrated either by experiments from our laboratory or from descriptions in the literature.

1. A compound with primary amino and carboxylic acid functions p-amino-benzoic acid

This was introduced in a solution of 50 % aqueous acetonitrile containing 0.01% formic acid, pH ca. 3.

Note the complementary positive and negative ESI spectra in which the protonated molecule is seen in positive ion mode and the negative ion, presumably the carboxylate anion, is observed in negative ion mode. Note too the commonly observed result that carboxylic acids can be detected with greatest sensitivity as anions even though the solution phase is at low pH. This has been referred to as 'wrong way round ionisation'.

Spectrum 1

Spectrum 1

2. Unsaturated lipids with oxygen functions

Many oxygen function containing compounds are amenable to ionisation in an ESI source. ab-unsaturated ketones and alcohols, sugar hydroxyl groups on glycosides, and many others can be analysed in an ESI source. Ando et al. have analysed a series of unsaturated hydrocarbon epoxides, constituents of insect sex pheremones. The model compound cyclohexene epoxide gives a good signal in positive ESI. Note the [M+H]+ parent ion and a fragment ion arising from loss of H2O.

Spectrum 2

Spectrum 2

3. Glycosides and similar metabolites

Phenolic compounds ESI is an important tool for the characterisation of naturally occurring phenolic compounds such as flavanoids and related glycosides. Lamuela-Raventos et. al. have described their analysis from acetonitrile and aq. formic acid mixtures. The spectra below show the results of the analysis of naringin by ESI together with the response from an ApCI source.

Spectrum 3


Spectrum 3

Spectrum 4

Spectrum 4

Spectrum 5

Spectrum 5

Spectrum 3 shows the positive and spectrum 4, negative ESI results for naringin. Note the sodiated adduct in the positive ESI experiment, the ion at m/z 581 being the [M+H]+ species. This adduction is absent in the negative ion spectrum. The ion at m/z 273 is the aglycone. Useful fragmentation is observed in the positive spectrum. Spectrum 5 is the negative ion ApCI spectrum of the same species. Much more fragmentation is observed. The positive ion ApCI spectrum (not shown) is essentially identical to the positive ESI spectrum.

4. Covalent and non-covalent complexes with metal ions

Cationisation by metal ions is common observation in ESI spectra and the phenomenon can be of great use in understanding the structures of such complexes. Moriwaki H, has examined the complexing of Cadmium ions with guanine bases in an attempt to understand the mechanism of Cadmium toxicity and its action on DNA. The spectra below show one such spectrum obtained by spraying a mixture of deoxyguanoisine and Cadmium acetate (10:1 molar proportions) into a positive ion ESI source.

Spectrum 6

Spectrum 6

Note in Spectrum 6, the extensive family of ions each representing a complex adduct of one or more base molecules with one or more Cadmium ions. The spectrum also contains protonated deoxyguanosine adducts.

Spectra 7 and 8

Spectra 7 and 8

Spectra 7 and 8 show two regions of the full spectrum. The Spectrum 7 shows an ion arising from molecules of base complexing with one cadmium ion. This ion carries two charges, as is evident from the 0.5 dalton spacing of the typical cadmium isotope pattern, this ion represents the complex [(deoxyguanosine)3Cd]++. Spectrum 8 shows a deoxyguanosine containing protonated ion carrying a single charge; here the isotope spacing is 1 dalton, there is no sign of the Cadmium isotope pattern and the m/z corresponds to the ion [(deoxyguanosine)2Na]+.

5. PAH molecules with oxygen functions

ESI is not the best method for the detection of PAHs. However if they contain a phenolic hydroxyl as in the case of Phenanthrol good spectra can be obtained by using ApCI in either polarity. The figure below shows the result of such an analysis with clear [M+H]+ (lower panel) and [M-H]- ions (upper panel) forming a complementary pair. The sample was run in a methanol:water:acetone solvent mixture.

Spectrum 9

Spectrum 9

6. Aromatic hydrocarbons

The analysis of the parent PAH hydrocarbon is more difficult; ESI and ApCI on the pure material do not provide satisfactory results; however if a solution is analysed containing a small concentration of silver ions in ESI clear results can be obtained. In positive ion mode the attachment of one Ag+ ion leads to an [M+Ag]+ ion at m/z 285/287 and also a pair of ions at 317/319 which are extra methanol attachment ions arising from the solvent, Spectrum 10. If a high cone voltage is applied in-source fragmentation leads to the free radical ion M+° at m/z 178, Spectrum 11. This gives a spectrum comparable to the electron ionisation spectrum of phenanthrene.

Spectrum 10

Spectrum 10

Spectrum 11

Spectrum 11

7. Intractable Nitrogen compounds which are Electrophilic

A nitrogen containing molecule C4HN2Cl3 which gave no signal in ESI but could be analysed by negative ion ApCI in electron capture mode. Spectrum 12 below shows the formation of the expected M-° chlorine isotope family at m/z 182/186 and a signal at m/z 163/165 for a two chlorine atom containing impurity arising from the synthesis.

Spectrum 12

Spectrum 12

8. High sensitivity detection of lipid acids

This electron capture ApCI source has been used for quantitative analysis of derivatised lipid carboxylic acids at the limits of detection. The process involves the derivatisation of the acid to an electron capturing molecule by esterification to a pentafluorobenzyl ester. This ester initially captures an electron to form an [M-°] radical anion which immediately loses a pentafluorobenzyl radical leaving the original acid as the charged anion R-COO-. The spectrum of 9,10-dihydroxystearic acid is shown below.

Spectrum 13


Spectrum 13

9. Terpenoids

Non-polar terpenoids are harder to ionise. Those which contain an ab-unsaturated hydroxyl group, such as testosterone, will protonate to give abundant [M+H]+ signals but a simple triterpene such as cholesterol does not respond to ESI ionisation. The literature contains references to the formation of chlorine attachment ions in negative ion ESI i.e. [M+Cl]- by spraying the analyte from a solvent such as chloroform as well as several accounts of successful ionisation by APPI. Alternatively good signals can be obtained in positive ApCI where no molecular ion is formed but a water loss fragment is detected. The spectrum below shows the formation of the [M-H2O]+ ion from cholesterol.

Spectrum 14

Spectrum 14

Inorganic ions are easily detected but can form clusters of attachment ions. The spectrum below is that from a mixture of NaI and CsI analysed from an aqueous solvent. The spectrum below is employed to provide signals for the calibration of the m/z scale in many instruments.

Spectrum 15

Spectrum 15

Modifications to Molecules to Accept Ionisation by ESI or Increase Response

The excellent response available from ESI sources has prompted many workers to adopt a strategy whereby they modify the structure of the analyte molecule to promote it ionisation in an ESI source. This can be achieved either by introducing a structure which already carries a charge or by a structure which is strongly basic or acidic or by one which will easily be oxidised leading to the formation of a radical cation.

There is a wealth of references to such procedures in the literature and I have given below only a few to illustrate the strategy.

1. Enhanced detection of naturally occurring steroids containing aldehyde and ketone functions have been achieved by reaction with the ionised reagent Girard T. This reacts with a ketone group to give a derivative containing a quaternary ammonium group as in the diagram below:

Derivitization with Girard's T reagent

Derivitization with Girard's T reagent

Griffiths et al. have extended this process with a variety of similar reagents and have optimised not only the ESI response but the tandem mass spectrometric behaviour as well.

2. David Harvey has developed a range of derivatisation techniques appropriate for the detection and structural characterisation of carbohydrates and glycosylated molecules. One strategy has been to ionise in the presence of common inorganic anions such as nitrate. Strong signals of adducts were observed in negative ion ESI. An alternative methods has used 2-amino-benzoic acids as derivatising agents for N-linked glycans which then give prominent [M-H]- ions.

3. We have developed methods designed to perform quantitative analysis of naturally occurring lipid aldehydes relevant to oxidative stress in human vascular disease. The lipid aldehydes were derivatised with a reagent developed for fluorometric analysis several years ago in which cyclohexanedione and ammonia react with the aldehyde group to produce a nitrogen containing derivative. This is basic and can be ionised easily in an ESI source to give an [M+H]+ ion.

4. Williams et. al. has published a method for the derivatisation of carbohydrates with ferrocenyl boronate. This leads to a ferrocene containing molecule. In an ESI source modified to promote the oxidation of analytes in the flowing solvent stream a Ferrocenyl radical cation is formed the structure of which is shown below:

Ferrocenyl Boronate derivative

Ferrocenyl Boronate derivative

van Berkel have published a number of papers relating to the electrochemical aspects of electrospray mass spectrometry.

5. Cooks et. al. have made use of the non-covalent adducts that many substances form with metal ions and other substances to perform chiral analysis. A chiral amino acid and a metal ion, copper is commonly used will form complexes with chiral mixtures the two steroismeric species of which will be distinguishable in tandem mass spectrometry in that they give different intensities of their product ions according to the original steroisomer present. This provides a method for steroisomeric analysis on small quantities of material where insufficient is present for a full NMR determination to take place.


There is wealth of reported experiments in which small molecules are being examined by API methods. Probably the greatest use of API techniques is employed by the pharmaceutical industry in qualitative and quantitative studies related to drug development. The emphasis here is on rapid automated multisample processing. A recent review of the applications to metabolic studies22 reveals the extent of experimentation in this field. Apart from the examples mentioned above, analysis of inorganic ions, organo-metallic species, mechanisms of reaction and thermodynamic parameters are also being investigated. It is important to have some understanding of the mechanisms of ionisation in the various atmospheric pressure systems if the optimum conditions for applying the sample are to be obtained. As our understanding of the basic mechanisms involved in ESI grows I am sure new and better procedures will continue to be developed.  


1. J.B. Fenn, M Mann, C.K. Meng, S.F. Wong and C.M. Whitehouse. Science, 1989: 246; 64.

2. Nature 1994:367;21.

3. Tong, J. Am. Soc. Mass Spectrom. 1999: 10; 1174-87.

4. T.C. Rohner, N. Lion, H.H. Girault. Phys. Chem. Phys. Chem. 2004: 16; 3056-3068.

5. T.R. Covey, A.P. Bruins. Anal. Chem. 2000: 72; 3653-59. and T.J. Kaupola, R. Kostiainen, A.P. Bruins. Rapid Commun. Mass Spectrom. 2004: 18; 808.

6. F.A. Straube, W. Dekant, W. Volkel, J. Am. Soc. Mass Spectrom. 2004: 15; 1853-62.

7. JM Purcell, CL Hendrickson, RP Rogers, AG marshall. Anal. Chem., 2006:78 5906-5912.

8. Z Takats, JM Wiseman, B Gologan, RG Cooks. Science 2004: 306; 471-473. RG Cooks, 54th ASMS Conference on Mass Spectrometry, Seattle, Wa. June 2006. poster

9. T Ando J. Mass Spectrom. 2003: 38; 328-332.

10. F. Sanchez-Rabanedu J. Mass Spectrom. 2003: 38; 35-42

11. H. Moriwaki. J. Mass Spectrom. 2003: 38; 321-327.

12. X. Xu, J. Zhng, L. Zhang, W. Liu, C.P. Weisel. Rapid commun. Mass spectrom. 2004: 18; 2299-2308.

13. K.M. Ng, N.L. Ma, C.W. Tsang. Rapid Commun. Mass Spectrom. 2003: 117; 2082-88.

14. G. Sing, A Gutierrez, K. Xu, I.A. Blair. Anal. Chem. 2000: 72; 3007-13.

15. R.B. Cole, J Zhu. Rapid Commun. Mass Spectrom. 1999: 13; 607-11.

16. J.J. Palmgrén, A Toyras, T Mauriala, J Monkkonen, S Auriola. J. Chrom B. 2005: 821; 144-152.

17. W.J. Griffiths. Mass Spectrom. Rev. 2003: 23; 81-152

18. D.J. Harvey. J. Am. Soc. Mass Spectrom. 2005: 16; 622-630, 631- 646, 647-659.

19. D.J Harvey. Rapid Commun. Mass Spectrom. 2005:19; 397-400.

20. I O'Brien-Coker, G Perkins, A I Mallet. Rapid Comm. Mass Spectrom. 2001: 15; 920-928.

21. M.K. Young, N. Dinh, D. Williams. Rapid Commun. Mass Spectrom. 2000: 14 1462-67.

22. G.L. van Berkel 1991: 63; 2064 and 1998: 70; 1544-1554.

23. W.A. Tao, F.C. Gozzo, R.C. Cooks. Anal. Chem. 2001: 73; 1692-98.

24. S Auriola J. Mass Spectrom. 2003: 38; 357-372. 

Social Links

Share This Links

Bookmark and Share


Suppliers Selection
Societies Selection

Banner Ad

Click here to see
all job opportunities

Most Viewed

Copyright Information

Interested in separation science? Visit our sister site

Copyright © 2017 John Wiley & Sons, Inc. All Rights Reserved