electron ionization lc-ms with supersonic molecular beams ...polaris/papers/2015 - ei-lc-ms...

Research article

Journal of

MASS SPECTROMETRY

Received: 5 July 2015 Revised: 19 August 2015 Accepted: 20 August 2015 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jms.3695

1252

Electron ionization LC-MS with supersonicmolecular beams—the new concept, benefitsand applicationsBoaz Seemann, Tal Alon, Svetlana Tsizin, Alexander B. Fialkovand Aviv Amirav*

A new type of electron ionization LC-MS with supersonic molecular beams (EI-LC-MS with SMB) is described. This system and itsoperational methods are based on pneumatic spray formation of the LC liquid flow in a heated spray vaporization chamber, fullsample thermal vaporization and subsequent electron ionization of vibrationally cold molecules in supersonic molecular beams.The vaporized sample compounds are transferred into a supersonic nozzle via a flow restrictor capillary. Consequently, while thepneumatic spray is formed and vaporized at above atmospheric pressure the supersonic nozzle backing pressure is about 0.15Barfor the formation of supersonic molecular beams with vibrationally cold samplemolecules without cluster formation with the sol-vent vapor. The sample compounds are ionized in a fly-though EI ion source as vibrationally cold molecules in the SMB, resultingin ‘Cold EI’ (EI of vibrationally cold molecules) mass spectra that exhibit the standard EI fragments combined with enhanced mo-lecular ions. We evaluated the EI-LC-MS with SMB system and demonstrated its effectiveness in NIST library sample identificationwhich is complemented with the availability of enhancedmolecular ions. The EI-LC-MS with SMB system is characterized by linearresponse of five orders ofmagnitude and uniform compound independent response including for non-polar compounds. This fea-ture improves sample quantitation that can be approximated without compound specific calibration. Cold EI, like EI, is free fromion suppression and/or enhancement effects (that plague ESI and/or APCI) which facilitate faster LC separation because full sep-aration is not essential. The absence of ion suppression effects enables the exploration of fast flow injection MS-MS as an alterna-tive to lengthy LC-MS analysis. These features are demonstrated in a few examples, and the analysis of the main ingredients ofCannabis on a few Cannabis flower extracts is demonstrated. Finally, the advantages of EI-LC-MS with SMB are listed anddiscussed. Copyright © 2015 John Wiley & Sons, Ltd.

Keywords: supersonic molecular beams; LC-MS; electron ionization; Cold EI; cannabis flower analysis

* Correspondence to: Aviv Amirav, School of Chemistry, Tel Aviv University, Tel Aviv69978, Israel. E-mail: [email protected]

School of Chemistry, Tel Aviv University, Tel Aviv 69978, Israel

Introduction

Liquid chromatography–mass spectrometry (LC-MS) has becomean established, widely used technique. Electrospray ionization(ESI)[1–3] is by far the most widely used LC-MS ionization methodbecause of its superior sensitivity, robustness and extended samplemolecular weight range, enabled by the formation of multiplycharged ions. ESI is supplemented with atmospheric pressurechemical ionization (APCI)[4,5] and atmospheric pressure photo ion-ization (APPI),[6–8] which in some cases, show better performancewith relatively small and less polar compounds. However, ESI, APCIand APPI still suffer from limitations in the ionization of non-polarcompounds, are characterized by non-uniform compound-specificresponse and are plagued by ion suppression effects. In addition, allthese atmospheric pressure ionization (API) methods are soft tech-niques that usually produce protonated or deprotonatedmolecularion (with or without adducts), hence requiring MS-MS or high reso-lution accurate mass MS for analyte identification and characteriza-tion. As a result, they are expensive and do not share the benefit ofgas chromatography–mass spectrometry (GC-MS) of having 70-eVelectron ionization (EI) mass spectra with library based sampleidentification that uniquely provide automated identification withnames and structures at the isomer level. Furthermore EI providesinformation rich fragmentation pattern with significant amount ofstructural information.[9]

J. Mass Spectrom. 2015, 50, 1252–1263

The main EI-LC-MS approach that was used until about a decadeago, particle-beam (PB) LC-MS,[10–15] suffers from limited sensitivity,nonlinear response and adverse matrix effects. Cappiello and co-workers demonstrated significant advances in the developmentof their unique capillary scale PB-LC-MS[16–19] and in the last 13years in their advanced direct EI[20–32] which excels with low LCflow rates while producing library searchable classical EI massspectra. Moreover, Cappiello and co-workers gave compellingevidence for the lack of matrix related ion suppression or en-hancement effects in their direct EI LC-MS[24,30] and demonstratedits strength and usefulness in a few selected interestingapplications.[25,27,29,32]

While standard 70-eV EI is a powerful ionization method forunknown sample identification, it is not ideal. About 30% of theNIST library compounds either have a weak (below 2% relativeabundance) or no molecular ion. This problem is further exacer-bated for LC related compounds which are larger and morethermally labile than standard GC-MS compounds. Furthermore,LC-MS compounds are usually less volatile and therefore require

Copyright © 2015 John Wiley & Sons, Ltd.

Electron ionization LC-MS

Journal of

MASS SPECTROMETRY

125

higher standard EI ion-source temperatures with the conse-quence of further intra ion-source degradation and lower mo-lecular ion relative abundance. Without a molecular ion, EIbased sample identification is not as trustworthy. Furthermore,for compounds that are not included in the EI libraries the ab-sence of the molecular ion severely hampers the usability ofthe EI mass spectra. Thus, the ‘ideal’ ionization method shouldprovide the informative, library searchable EI fragments com-bined with an enhanced molecular ion (relative to standardthermal EI), whose observation as the largest m/z mass spectralpeak should be trusted.[33]

We developed and explored the interface of GC and MS withsupersonic molecular beams (SMB) and the use of a fly-throughEI ion source as a medium for EI of vibrationally cold samplecompounds in the SMB (hence also named GC-MS with ColdEI). We found that GC-MS with Cold EI provides enhancedmolecular ions, extended range of compounds amenable forGC-MS analysis, improved sensitivity particularly for com-pounds that are difficult to analyze, facilitates faster GC-MSanalysis and provides relatively uniform compound indepen-dent ionization yields. GC-MS with SMB (Cold EI) is reviewed inreference.[33,34]

We also explored electron ionization LC-MS and in our pastpublications[35–39] we described a new method and instrumenta-tion for the combination of LC-MS and EI through the use of super-sonic molecular beams (SMB) as a medium for EI of vibrationallycold sample compounds (hence also named Cold EI). The reportednew (at that time) approach of EI-LC-MS with SMB[35–37] was basedon spray formation behind a supersonic nozzle, followed by fullthermal sample particle vaporization prior to sample expansion asisolated molecules from the supersonic nozzle. The supersonic freejet was collimated and formed a supersonic molecular beam thatcontains vibrationally cold sample molecules. These molecules pro-ceed axially along a dual cage fly-through EI ion-source[40] forobtaining Cold EI mass spectra with enhancedmolecular ions. Clus-ter formation with the solvent vapor was practically eliminated byusing a relatively large diameter and separately temperature-controlled nozzle. Vaporized solvent molecules served as the SMBcarrier gas without adding another seeding gas. LC-MS with SMBwas experimentally evaluated for a broad range of thermally labileand relatively large compounds in liquids,[35–37] and it generatedthe following main advantages: (1) Library searchable EI mass spec-tra are provided for positive sample compounds identification,combined with enhanced molecular ions for improved confidencelevel in the identification; (2) Non-polar compounds are amenablefor analysis in addition to the standard range of APCI compounds;(3) Approximately uniform compound independent response wasobtained which enables easier sample quantitation even for un-known samples. (4) Matrix related ion suppression effects do notexist.

However, our past approach and system suffered from a majorproblem of poor robustness as the use of thermally assisted spray(Thermospray, also sometimes known as ‘spray and pray’) resultedin frequent clogging of the solvent delivery tube near its exit be-cause of sample degradation and solvent salt residue deposits inthat heated place (like limestone deposits in domestic kettles). Fur-thermore, service to replace the clogged solvent delivery tube re-quired full venting of the system and thus was tedious and timeconsuming. In addition, because all the LC solvent was exhaustedinto the nozzle and vacuum system the liquid solvent flow ratewas limited to practically 50μl/min depending on the vacuumpump flow rate acceptance. Thus, our research was continued to

J. Mass Spectrom. 2015, 50, 1252–1263 Copyright © 2015 John

improve our EI-LC-MS with SMB approach in order to retain itsbenefits while eliminating its abovementioned deficiencies. Con-sequently, the main challenge in having EI-LC-MS with SMB ishow to achieve a reliable and robust conversion of steady LC out-put liquid flow into gas phase (vapor) flow with the followingfeatures:

a) No cluster formation.b) No nozzle and/or solvent delivery tube clogging.c) In a robust and easy to maintain system.d) For broad range of compounds.e) With a stable spray.f) For useful solvent flow rate range of 20–250μl/min.g) With acceptable helium nebulizing flow rate.h) With minimal added cost.i) Dual use GC-MS and LC-MS with SMB (Cold EI) in one system.

In this paper we describe our new EI-LC-MS with SMB methodand system while describing how it retains the past demonstratedbenefits together with the full elimination of its past downsides toprovide an effective and useful new type of LC-MS.

Experimental—the EI-LC-MS with SMB system

In order to design an optimal EI-LC-MS systemwe started with hav-ing a strategy and list of goals for instrument development. Accord-ingly, our system and its components are based on the followingnovel concepts and considerations:

1. Use of a supersonic molecular beams (SMB) interface and itsfly-through EI ion source in order to obtain enhancedmolec-ular ions and extended range of compounds amenable foranalysis. This range should be bigger than that of the parti-cle beams LC-MS or GC-MS (equivalent to that of APCI plusthe non-polar compounds).

2. Use of a quadrupole MS for lower cost because the ability toidentify sample compounds with the library implies thathigh cost high resolution MS is not essential.

3. Use exactly the same SMB interface and fly-through ionsource as in GC-MS with SMB in order to have the optionfor both GC-MS with Cold EI and EI-LC-MS with SMB in thesame system with an easy interchange procedure.

4. Use of triple quadrupole MS for the experimental system tohave MS-MS capability in order to explore the option to an-alyze samples in complex matrices via much faster flowinjection MS-MS instead of lengthy LC-MS (in view of ab-sence of ion suppression effects).

5. Employ high pressure nebulization (1–1.5 Bar) as in APCI forimproved robustness and spray reliability as well as for easyvaporization chamber service without venting.

6. Perform the spray vaporization inside a standard GC liner foreasy and low cost replacement service in case of linercontamination.

7. Connect the vaporization chamber to the SMB nozzle with ashort heated deactivated fused silica capillary transfer line(neutral funnel) in order to reduce the nozzle backing pres-sure to suppress and eliminate sample cluster formationwith the solvent vapor.

8. Use of a larger diameter nozzle (0.3-mm ID instead of0.1-mm ID) for lower pressure supersonic jet expansion tosuppress cluster formation while retaining the SMB vibra-tional cooling.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jms

3

B. Seemann et al.

Journal of

MASS SPECTROMETRY

1254

9. Design the heated sample vaporization chamber with maxi-mumaxial temperaturegradientat itsentrysideforbestspraystability (eliminate onset Thermospray spray instabilities).

10. Use helium as the nebulization gas for improved nozzle-skimmer jet separation and higher signal.

11. Develop an optimal pneumatic spray head that will producestable spray at wide range of solvent flow rates and compo-sition at minimal helium flow rate.

12. Develop a vaporization chamber split valve that will enablethe use of pneumatic spray with increased LC solvent flowrates given the limit of total (helium and solvent) flow rateacceptance of the turbo molecular pump of the SMB inter-face vacuum chamber.

13. Design a spray head axial positioning device for its optimaltemperature and spray stability positioning inside the hea-ted vaporization liner.

14. Design an axial positioning device for the solvent deliverytube for its placement at the optimal position inside thepneumatic spray head.

15. Design the whole EI-LC-MS spray vaporization device as aone unit from solvent delivery tube up to the supersonic noz-zle in order to enable GC-MS with Cold EI and EI-LC-MS in aone system that can be interchanged via the change of thevaporization chamber unit with the GC-MS transfer line only.

Having the above design considerations in mind we shall nowdescribe the experimental EI-LC-MS system with some emphasison how it achieves these goals. The description will begin withthe SMB-interface, ion source and MS system and after that theEI-LC-MS vaporization chamber and interface will be described.The EI-LC-MS with SMB apparatus is schematically shown in Fig. 1.It is based on our modified home-made GC-MS with SMB (GC-MSwith Cold EI) system that was previously described in details.[34]

The base MS system is a Varian 1200 L triple quadrupole MS systemthat was modified to include the SMB interface and fly-through ionsource, and we named it 1200-SMB.In the 1200-SMB the supersonic nozzle expands 60–90ml/min

mixture of vaporized solvent, vaporized sample compounds andheliumnebulization andmake up gas into a nozzle (SMB formation)vacuum chamber that is differentially pumped by a Varian Navigator

Figure 1. Electron ionization LC-MS with SMB system outline. The liquid is intheated vaporization chamber through a pneumatic nebulizer. The helium nsheath gas line and nozzle make up gas line (indicated). The SMB interface issample molecules fly through the ion source while the produced ions followed

wileyonlinelibrary.com/journal/jms Copyright © 2015 Joh

301 turbo molecular pump (Varian Inc., Torino Italy) with 250 l/spumping speed. The pressure at this vacuum chamber is about6μBar. While most of the helium flow arrives from the pneumaticnebulizer some low helium flow rate (~5ml/min) can be separatelyprovided directly to the nozzle, and it can be mixed (via an openingof one valve) with perfluorotributylamine (PFTBA) for periodic sys-tem tuning and calibration. The sample compounds seeded in thehelium and solvent vapor expand from a 300-μm diameter super-sonic nozzle. The supersonic expansion vibrationally cools the sam-ple compounds, and the expanded supersonic free jet is skimmedby a 0.8mm skimmer and collimated in a second differentiallypumped vacuum chamber, where an SMB is formed. The secondvacuum chamber is pumped by a Varian 400/300 split turbo molec-ular pump that pumps both the second vacuum chamber (400 l/s)and main MS vacuum chamber (300 l/s). The SMB seeded withvibrationally cold sample compounds flies through a dual cage EIion source[40] where these beam species are ionized by 70-eV elec-trons with 1–2-mA emission current. The ions are focused by anion lens system, deflected 90° by an ionmirror and enter a radio fre-quency (RF)-only hexapole ion transfer optics (Q0 of the originalVarian 1200 system). The 90° ion mirror is separately heated andserves to keep the mass analyzers clean from sample induced con-taminations. The ions are further transferred through an ion lensinto the MS vacuum chamber and are analyzed by a quadrupoleMS-MSmass analyzer system. It consists of two quadrupolemass an-alyzers (Q1 and Q3) and a collision cell (Q2). As in any quadrupoleMS-MS system, it can operate in SIM or full scan mode, as well asin all common MS-MS scan modes. Because Q2 is a 180° curvedRF-only quadrupole ion transfer system in the 1200 L, a head-onion detector is positioned directly in the path of ions exiting Q3.Its entrance is biased at 5 KV, serving as an efficient ion to electronconverter. In this configuration we already achieve several goalsfrom the above list of 15 goals. The SMB interface and its fly throughion source provide Cold EI mass spectra with enhanced molecularions, and this is the same system as used in GC-MS with Cold EI.The use of big nozzle with 300-μm ID serves to reduce its opera-tional pressure hence suppress and practically eliminate samplecluster formation with the solvent vapor.

In Fig. 2 we show the EI-LC-MS sample vaporization chamber andits interface with the supersonic nozzle. Going from left to right in

roduced either from HPLC system after its column or a syringe pump to theebulization gas enter the SMB interface through a nebulization gas line,separately pumped by a turbo molecular pump and the vibrationally cold90° deflection into the Varian 1200 L triple quadrupole MS.

n Wiley & Sons, Ltd. J. Mass Spectrom. 2015, 50, 1252–1263

Figure 2. Detailed view of the EI-LC-MS interface, from the liquid inlet to the skimmer. The adjustable spray nebulizer is located between the cooling unitand the heater block for maximum axial temperature gradient. The center of the interface is the separately heated sample spray formation and vaporization.The separately heated transfer-line also accommodates the fused silica capillary that serves as a flow restrictor which governs the gas flow rate into the SMBnozzle and vacuum chamber. The supersonic molecular beam nozzle center is adjusted to the skimmer concentrically by the X–Y–Z tune.


Journal of

MASS SPECTROMETRY

125

Fig. 2 we shall now describe the EI-LC-MS vaporization chamberalong the liquid flow path. Samples were introduced using a HPLC(model 1100, Agilent Technologies, Waldbroon Germany) (or mo-del ProStar LC, Varian Inc, Walnut Creek CA).

In some cases samples were injected using a 5-μl injection loop,directly into a liquid transfer line without a separation column. Theliquid transfer line was typically made of fused silica capillary with75-μm ID and 190-μm OD provided by PolyMicro Technologies(Phoenix Arizona, USA). Methanol and water were used as the sol-vents of choice at flow rates in the 20–250μl/min range while ace-tonitrile, hexane, iso-propanol, ethylacetate and other solventswere also used. Acetonitrile required the use of special rhenium fil-ament wire as it etched the tungsten ribbon filaments. The solventflow in the capillary entered the EI-LC-MS system through the Z-Axis capillary tune device that enabled the careful positioning ofthe solvent delivery capillary at the pneumatic spray nozzle. Wefound that axial positioning within 0.2mm from its optimum placeis desirable for having the most stable spray for the broadest sol-vent flow rates range and with relatively low helium nebulizinggas flow rate. The solvent delivery capillary further entered the liq-uid nebulization chamber via a Z-Axis spray probe tune device thatcan mechanically adjust the position of the pneumatic spray nozzlefrom 6mmbefore the entrance to the heated vaporization oven upto 20mm inside the vaporization oven. The purpose of this Z Axistuning device is to optimize the position of the spray head in a lo-cation that is not yet hot, in order to suppress onset of Thermosprayrelated spray instabilities, while emitting spray that touches theheated walls of the vaporization chamber at surface points whichare sufficiently hot to vaporize the sample without any peak tailing.As a result, the spray formation and vaporization chamber is de-signed according to the concept of maximum axial temperaturegradient. A glass liner serves as the heated vaporization chamber,and it is air cooled at its entrance by aluminum cooling block(Cooling unit), that despite being close within 10mm from theheated vaporization chamber that is maintained at typically300 °C it is at about 40 °C only. The glass liner (standard GC glassliner) which is a very good thermal insulator directs the NozzleSpray head into the heated vaporization chamber. As a result, be-cause of themaximumaxial temperature gradient the liquid outputfrom the LC is retained as liquid without premature formation ofbubbles which mark the onset of Thermospray and produce unde-sirable spray instability. Typical spray nozzle head location is about


6mm inside the heated vaporization chamber. The spray nozzlehead has a unique design that is based on our modification ofthe concept of ‘flow focusing’ nebulizer.[16,41] In this spray nozzlethe solvent delivery capillary is centered at a narrow neck that alsoenables the flow of nebulizing helium gas. The solvent delivery cap-illary ends at about 0.1mm from the exit of the spray head tubewith 0.5-mm ID that ends with 0.25-mm ID cup. We found that thisstructure provides a stable spray that is directed forward at a nar-row angle so that the spray touches the surface of the heated va-porization chamber in a place where it is fully heated to its settemperature. The vaporization chamber is based on a standardGC liner which is a deactivated glass. We use a liner of Agilent7890 GC with 78-mm length, 6.35-mm OD and 4-mm ID. We foundthat the Jennings cup liner seems to be the most suitable liner as itdoes not require glass wool that can exhibit surface activity yet itprevents the penetration of unvaporized particles into the capillarytransfer line and ensures full vaporization at the inert glass liner. Thevaporization chamber is separately heated and temperature con-trolled. The liquid solvent is pneumatically transformed into a spray,the sprayed solvent is vaporized and the emerging sample particlesare fully vaporized, similarly to their vaporization in APCI or APPI.Helium is used for nebulization at typically 150ml/min from which130ml/min serves for nebulization and about 20ml/min as a sheathgas to prevent back migration and tailing of sample compounds.One electronic flow control stabilizes the helium flow rate whichis split with flow restrictors into the flow rates as above. The heatedsample vaporization chamber also includes a split out tube andvalve for the gas after the liner exit, in similarity to the split gas linein GC injectors. Consequently, the vaporization chamber pressure ismechanically adjusted via a valve knob to provide about 45–60Wturbo pump power consumption which emerges from flow rateof about 60–90ml/min. This split valve allows the use of sufficienthelium flow rate as needed for achieving stable pneumatic sprayas well as liquid flow rate up to 250μl/min without overloadingthe turbo pump. Typical operation is at 1.5 Bar with split ratio of1:1.5. After the sample is fully vaporized at the heated liner it isswept with the helium nebulizing gas and solvent vapor into a sep-arately heated and temperature controlled deactivated fused silicacapillary transfer line with 0.25mm ID. This capillary that separatesthe high pressure (1–1.5 Bar) vaporization chamber and the nozzlewhich is maintained at about 0.15Bar via its large diameter (0.3mmID) serves to optimize the sample vaporization process and enables


5

B. Seemann et al.

Journal of

MASS SPECTROMETRY

1256

its service at high pressure yet to maintain the nozzle at its optimalbacking pressure to provide good vibrational cooling without sam-ple and solvent vapor cluster formation.[42] Thus, the transfer linecapillary serves as flow impedance between the vaporization cham-ber and nozzle. At the length of 22 cm and at 300 °C it deliversabout 28ml/min helium flow rate at ambient 1 Bar pressure in thevaporization chamber (open split valve). The whole structure of sol-vent delivery capillary and its axial adjustments devices, spray noz-zle, heated vaporization chamber and its split line, heated transferline and nozzle are mounted as a one unit into an XYZ adjustmentmechanism on the nozzle vacuum chamber. Thus, the nozzle canbe accurately positioned in front of the skimmer at the optimal dis-tance of about 12mm from the skimmer. The dual cage fly-throughion source is placed a few millimeters after the skimmer. As ex-plained above, with this design all our 15 goals as described atthe beginning of this chapter are achieved.The nozzle diameter is an important parameter that is linked to

the useful liquid flow rate range. It must be sufficiently large to elim-inate post expansion cluster formation,[42] yet small enough to en-able effective vibrational cooling. Cluster formation of the samplewith the solventmolecules depends on three body collisions, henceon P3D, while the vibrational cooling depends on PD, where P is thepressure behind the nozzle and D is the nozzle diameter. Thus, for agiven liquid solvent flow rate, the tripling of the nozzle diameter(from its GC with Cold EI diameter) leads to the reduction of the va-porized solvent pressure by a factor of 9 hence to the reduction ofcluster formation by a factor of 729 (=93). Meanwhile, such triplingof the nozzle diameter has only a minor penalty factor of 3 on theefficiency of vibrational cooling, that is far superior to that of purehelium in view ofmethanol or water being a heavier molecule with-out velocity slip.[43,44]

Initially, we assumed that achieving the fastest possible samplevaporization would lead to the softest sample vaporization be-cause of minimizing the time during which the sample can de-grade. However, in view of our more recent experience with theanalysis of thermally labile compounds in GC-MS, we came tothe opposite conclusion that it is better to employ slower vapori-zation at cooler conditions because temperature has a greater ef-fect than time on the promotion of thermal degradation.[45] As aresult, we now feel that the use of larger diameter glass liner with4mm ID as a vaporization chamber is preferable, and certainly it ismore robust in terms of requiring less frequent internal cleaning.We note that as a result, in comparison with typical APCI or APPIour sample thermal vaporization is slower yet performed at lowertemperatures such as 300 °C instead of 400 °C. From our experi-ence our conditions leads to softer thermal vaporization than inmost current APCI vaporizers.

Results—cold EI mass spectra, features andselected applications

In Fig. 3, the Cold EI mass spectrum of Trinitrotoluene (TNT) in me-thanol solution is shown in the left trace, and compared with itsNIST 08 EI library mass spectrum shown in the lower left trace. Notethe similarity of the library mass spectrum to the Cold EI MS ob-tained with the EI-LC-MS with SMB apparatus. All the major highmass fragment ions of m/z 210, 193, 179, 164, 149 and 134 as wellas the lower mass ions are exhibited in Cold EI the same as instandard EI, and thus good library search results are enabled withNIST library matching factor of 846, reversed matching factor of859 and 89.5% identification probability in the TNT identification.


In addition, the molecular ion at m/z=227 is now clearly observedwith 10% relative abundance while it is either weak or missing inthe library (2% relative abundance in one library MS and missingin several replica mass spectra). We also analyzed TNT in a standardAgilent 5977 MSD, and at 250 °C extractor ion source temperaturewe measured the relative abundance of the molecular ion to be0.05% andwith nearbym/z=228with 0.15% relative abundance be-cause of residual intra ion source self-chemical-ionization. The lowerGC-MS molecular ion relative abundance than one of the NIST li-brary MS is becausemany of the library mass spectra were obtainedwith low ion source temperatures that are not optimal for GC-MS us-age. In contrast, the molecular ion relative abundance with Cold EI isabout 10%.

While this 10% relative abundance of the molecular ion is nothigh it represents about 200 times increased relative molecularion abundance versus standard EI which now can be trusted andserve to verify the NIST library identification. We note that in ColdEI the NIST library search matching factors are high but somewhatlower than in standard EI for two reasons: (1) the molecular ionrelative abundance is increased; (2) the low mass fragment ionsrelative abundance is decreased. However, the identification prob-ability is high and often higher than in standard EI (as found in theanalysis of many pesticides[46]) because the molecular ion is themost important and sample characteristic ion in themass spectrumand the availability of themolecular ion serves well to verify the realidentity of the compound against other homologous and degrada-tion compounds. As shown, the overall effect of vibrational coolingon the appearance of the Cold EI mass spectrum is of shifting inten-sity from lowmass fragments to high mass fragments and in partic-ular to the molecular ion. In Fig. 3 we also show the Cold EI massspectrum of the Sulfamethoxazole drug (C10H11N3O3S) which iscompared with its NIST library in the middle two traces. This ther-mally labile drug is considered as not amenable for standardGC-MS analysis, but it is easy to analyze by the EI-LC-MS with SMBsystem. As shown the Sulfamethoxazole was identified as #1 inthe NIST hit list with 98.5% identification probability. As demon-strated the Cold EI mass spectrum shares with the NIST library massspectrum all the fragment ions and molecular ion while shifting in-tensity from low mass ions to the molecular ion whose relativeabundance is increased about four times. In Fig. 3 we further showat the right side the Cold EI mass spectrum of the Haloperidol drug(C21H23ClFNO2) that is comparedwith its NIST library (bottom right).This drug is also considered as not amenable for standard GC-MSanalysis but it is easy to analyze by the EI-LC-MS with SMB system.As shown the Haloperidol was identified as #1 in the NIST hit listwith 95.5% identification probability. As demonstrated, the ColdEI mass spectrum shares all the fragment ions with the NIST librarymass spectrum while exhibiting abundant molecular ion which ispractically missing in the NIST library mass spectrum (1–2% relativeabundance). Thus, Fig. 3 demonstrates the usefulness of the EI-LC-MS with SMB approach, in both the analysis of a thermally labilecompound (on the GC-MS scale) as well as in showing the benefitof mass spectra which show a combination of enhanced molecularion and standard EI fragments for correct library based sampleidentification. We note that while Cold EI enhances the molecularions the degree of enhancement is small for small molecules andthose with abundant or dominant molecular ions. On the otherhand, for largemolecules with high heat capacity the enhancementcan be significant and exceed a factor of 1000[47] so that large hy-drocarbons with <0.1% molecular ion relative abundances exhibitdominant molecular ions in Cold EI. We note that the availabilityof enhanced molecular ions is very important not only because it


Figure 3. Cold EI mass spectra of TNT, Sulfamethoxazole and Haloperidol as obtained with the EI-LC-MS with SMB system and its comparison with NISTlibrary classical EI mass spectra of these compounds. Flow injection of these compounds in methanol was employed with a 5-μl loop and methanol flowrate of 50 μl/min.


Journal of

MASS SPECTROMETRY

125

increases our confidence level in the correct sample identificationbut also because it enables the confirmation or rejection of NIST li-brary identification. Furthermore, while a quadrupole mass spec-trometer with unit resolution is used, elemental formula informationcan be obtained with the TAMI software that is based on isotopeabundance analysis of the molecular ion group of isotopomers[48]

which require the availability of abundantmolecular ions.An additional important beneficial feature of EI-LC-MS with SMB

is that its response is linear unlike that of the Particle Beam. In a pre-vious publication[36] we demonstrated over 4 orders of magnitudelinear dynamic range (LDR) for pyrene with linear correlation coef-ficient R= 0.9993 (as shown in Fig. 9 in that publication) and inour current system our LDR is five orders of magnitude from thelimit of detection which is about 1pg in SIM mode to about 1μgand the main limitation is in the proper adjustment of the ion de-tector gain within its limited dynamic range. We note that this highCold EI LDR is higher than that of typical ESI LDR.


Another important feature of the EI-LC-MS with SMB system isthat electron ionization in contrast to all the API ionizationmethods(ESI, APCI and APPI) is an in vacuum ionization method without anyion molecule reactions. Consequently and as demonstrated[28,31] EIdoes not exhibit ion suppression or enhancement effects thatplague the various API ionization methods. Cold EI, like EI, is an invacuum ionization method, and furthermore the ionization is per-formed in collision-free molecular beam environment; hence it iscompletely free from any ion suppression or enhancement effects.In Fig. 4 we demonstrate the absence of ion suppression orenhancement effects in the three consecutive flow injection analy-sis of 1 ppm Pyrene (SIM mode on its molecular ionm/z=202) thatare followed by three consecutive flow injection analysis of 1-ppmPyrene that was spiked with 1000ppm each caffeine (basic drug)and ibuprofen (acidic drug). As demonstrated, the 1000 timeshigher levels of added ‘matrix’ did not affect the Pyrene response.We note that Pyrene has negative proton affinity and is difficult


7

Figure 4. A demonstration that Cold EI is free of matrix ion suppressioneffects. Single ion mass chromatogram on the molecular ion of Pyrene (m/z= 202) is shown with three peaks that are the result of three flowinjections of 1 ppm Pyrene in methanol, followed by three peaks thatresult from the injection of 1 ppm Pyrene mixture with 0.1% caffeine and0.1% ibuprofen all in methanol.

Figure 5. Flow injection EI-MS-MS of Diazinon insecticide in a fruit andvegetable matrix is shown without ion suppression or enhancementeffects. The MS-MS signal of Diazinon is at its molecular ion m/z= 304 andproduct ion at m/z= 179. The first injection peak shows a fruit matrixspiked with Diazinon at 100 ppb (also 100 ng/g) concentration. The secondinjection peak is of the unspiked matrix, while the third injection peak is ofpure 100 ppb of Diazinon in acetonitrile. The liquid flow rate was set to15 μl/min acetonitrile. The measured signal to noise ratio is 80, and theestimate LOD is ~10 ppb based on the matrix equivalent signal.

B. Seemann et al.

Journal of

MASS SPECTROMETRY

1258

to ionize by ESI hence it is expected to be highly subjected to ionsuppression effects. Thus, while Fig. 4 provides only suggestive in-formation it can serve as an additional proof that Cold EI is free fromany ion suppression or enhancement effects.Because EI-LC-MS with SMB has no matrix related ion suppres-

sion effects, it does not always require LC for sample separationfrom the matrix, and thus a new approach is enabled for achievingultra-fast flow injection (FI) EI-MS-MS with SMB analysis. Recently,the topic of flow injection MS-MS as an alternative to lengthy LC-MS analysis received significant attention in the world of pesticidesanalysis in agricultural products.[49–53]

In these and in other experiments with electrospray ioniza-tion,[49–54] the sample extracts were diluted about 100–200 timesto reduce or eliminate ion suppression effects and analyzed by flowinjection, electrospray ionization MS-MS. However, as known, elec-trospray cannot serve for the ionization of all pesticides, especiallywith 100 times dilution, and its sensitivity is not sufficient for afew groups of pesticides, particularly those that currently requireGC-MS analysis such as the organo-chlorine pesticides. Our EI-LC-MS with SMB system can serve without any extract dilution, in itsflow injection (FI) mode, to supplement and complement ESI inthe development of fast and universal method for pesticide analy-sis. In Fig. 5, we show FI-EI-MS-MSwith SMB analysis of 100ppb (also100ng/g) diazinon in a QuEChERS mixed produce extract (providedby S.J Lehotay from the USDA). The first peak is of the sample spikedwith diazinon, the second peak is of the matrix alone and the thirdpeak is of 100ppb diazinon in pure acetonitrile. The matrix, whichcontained eight fruits and vegetables, gave several interferenceswith integrated intensity equivalent to about 10 ng/g diazinon. Asdemonstrated, the sum of the matrix signal [(Injection 2)] and diaz-inon in acetonitrile signal [(Injection3)] is the same as that of the di-azinon in matrix signal [(Injection 1)], thereby demonstrating theabsence of matrix related ion suppression effects. Thus, flow


injection EI-MS-MS with SMB could potentially serve for ultra-fastanalysis of sub ppm residues in agricultural matrices althoughmuchwork is required to properly evaluate this approach and test it withother pesticides and commodities.

In Fig. 6 we show the EI-LC-MS with SMB analysis of a mixture offour compounds (each at 20 ppm), Octafluoronaphthalene (OFN),Pyrene, Agidol 40 and Cholesterol. Isocratic analysis with 100%methanol was performed with 5-cm-long 1-mm ID LC column with3-μ C18 particles at 70μl/min solvent flow rate. A few features ofthe EI-LC-MS with SMB are demonstrated in Fig. 6: a) Fast isocraticLC-MS analysis is demonstrated with total analysis cycle time of2.5min. Because we are not concerned with ion suppression effectswe can tolerate partial co-elution and reduce the analysis time; b)All four compounds exhibit Cold EI mass spectra with dominantor abundantmolecular ions (including the large Agidol 40moleculewithMW=774) combinedwith the library searchable fragment ions;c) All these four compounds are relatively non-polar compoundsthat either cannot be ionized by ESI or APCI or are difficult to ana-lyze by standard ESI or APCI based LC-MS. d) All these four com-pounds were easily identified by the NIST library with highidentification probabilities as indicated in the inserts in the Fig. 6.In addition, when probabilities of the compounds’ isomers aresummed up the total identification probability (shown in Fig. 6) ishigher. e) The total ion count response was about uniform and var-ied in a range of only about 2–3 despite the large difference in theirstructures.

In Table 1 we show the relative response of these four com-pounds. The response of OFN and Cholesterol is close to a quartereach from the total while Pyrene is 1.7 higher and Agidol 40 is 2.5


Table 1. Relative abundance of peak areas of the four indicatedcompounds

RT [min] Area Percent

OFN 0.67 8.84E + 08 24%

Pyrene 0.842 1.54E + 09 42%

Agidol 40 1.489 3.60E + 08 10%

Cholesterol 2.067 8.66E + 08 24%

Figure 6. OFN, Pyrene, Agidol 40 and Cholesterol analysis by EI-LC-MS withSMB. Sample concentration was 20 ppm and C18 Phenomenex Lunacolumn was used with 3-μm particles, 50-mm length and 1.0-mm ID,using 5-μl sample loop. The LC method in this experiment was isocratic100% methanol at a flow rate of 70μl/min.

Figure 7. EI-LC-MS with SMB analysis of dimethylphthalate, ethylparabendiethylphthalate and butylparaben in order of their elution, at 50 ppmeach. The HPLC gradient used was from 50% water/methanol to 75%methanol with a solvent flow rate of 35 μl/min, in a 50mm×1.0mm C18Phenomenex Luna column with 3-μm particles.


Journal of

MASS SPECTROMETRY

125

times lower than expected based on uniform response. The lowerrelative response of Agidol 40 is attributed to the decline of thequadrupole mass spectrometer transmission at high masses. Wenote that the Pyrene EI mass spectrum shows almost no fragmentions below m/z=101 and thus scan start above m/z=100 (as inFig. 6) resulted in almost no fragments based TIC losses (unlike withother compounds); thus higher relative response was achieved.

This close to uniform response observation represents an impor-tant attribute of EI as an ionization method because in contrast toAPCI or ESI, its ionization efficiency is independent of the samplecompound polarity and identity. While the ionization yield is uni-form and compound independent the system response uniformityis somewhat eroded because of quadrupole (and its ion detector)


ion transmission decline with mass and the fact that low mass ionsbelow 70u are not collected. Furthermore, the response uniformityholds true only for pure compounds or simple mixtures. On theother hand, unlike in standard EI, in GC-MS with Cold EI no ionsource peak tailing affects the response uniformity. We estimatethat despite these imperfections the response uniformity allowsus to know (with a factor of 2) how much we have from each com-pound without the tedious and time consuming process of fullidentification, synthesis and the performance of compound specificcalibration curve.

In Fig. 7 the EI-LC-MS with SMB analysis of four simple com-poundsisshown.Dimethylphthalate,Diethylphthalate,Ethylparabenand Butylparaben were analyzed using gradient elution from 50%water/methanol to 75%methanol.

The upper trace shows the obtained total ion mass chromato-gram, while the bottom traces show the obtained Cold EI massspectra of diethylphthalate and ethylparaben and their respectiveNIST library mass spectra (inverted bottom MS). As demonstrated,very good similarities between the Cold EI and NIST library massspectra are visibly observed, and some enhanced molecular ionsare also shown. For diethylphtahalate the NIST library search


9

Figure 9. Sensitivity evaluation of the EI-LC-MS with SMB system in thesingle ion monitoring analysis of Cannabidiol (CBD) and δ9-THC analysis;40 ppb was injected through a 5-μl loop to a C18 Phenomenex Lunacolumn with 3-μm particles. The gradient was from 60% methanol and40% water to 100% methanol in 1.5min. The MS was monitoring atm/z= 314, and the Varian software calculated signal-to-noise ratio (peak topeak) is 293 for CBD and 737 for δ9-THC which translates to linearlyextrapolated LOD of 0.7 pg and 0.3 pg.

B. Seemann et al.

Journal of

MASS SPECTROMETRY

1260

matching factor was 883, reversed matching was 887 and identifi-cation probability was 77.6%. For ethylparaben the NIST librarysearch matching factor was 933, reversed matching was 940 andidentification probability was 80.4%. Thus, Fig. 7 demonstrates theapplicability of EI-LC-MS with SMB for the analysis of these com-pounds with effective library based identification. In addition,Fig. 7 also demonstrates the feature of response uniformity viathe similarity of the TIC peaks of all the four compounds despitehaving two different classes of compounds of paraben drugs andphthalate diesters.Cannabis and its main ingredients analysis became recently an

area of growing importance as while it is required for law enforce-ment it is also required for the optimization of its growth as toolfor product control in the increasing number of legal markets forcannabis plant products. Mass spectrometry and in particular EI-LC-MS with SMB can provide an effective and legally defensibleanalysis method for the detection and quantification of plant com-pounds and metabolites. In Fig. 8 the fast EI-LC-MS with SMB anal-ysis of the three main ingredients of Cannabis is shown (TIC at left)together with the obtained Cold EI and its identification results withthe NIST library.As shown, Cannabidiol (CBD), Cannabinol (CBN) and Tetrahydro-

cannabinol (THC) at 20 ppm are analyzed in under 2-min analysisand exhibit uniform compound independent TIC peak areas. TheCold EI mass spectra are highly characteristic and the Cold EI MSof CBD and THC are very different despite the fact that they are iso-mers that cannot be differentiated by ESI and/or MS-MS.[55] Asshown, the NIST library provided unambiguous identification withnames and structures at the isomer level with high (indicated) iden-tification probabilities.While these compounds are among the major components of

Cannabis and thus high sensitivity is not required in Cannabis anal-ysis, high sensitivity is always a feature of importance. Thus, whilewe evaluated the sensitivity of the EI-LC-MS with SMB system afew times in the past and found it to be with around 1pg LOD[36]

we decided to further evaluate it with additional compounds in

Figure 8. Cannabidiol (CBD), Cannabinol (CBN) and δ9-Tetrahydrocannabinonthrough a 5-μl loop into a C18 Phenomenex Luna column with 3-μm particles.methanol in water in 1min.


our new EI-LC-MS with SMB system. In Fig. 9 we demonstrate thesystem sensitivity in the single ion monitoring (SIM) analysis ofCBD and THC as in Fig. 8 but now at 40-ppb concentration levels(200pg on column with 5-μl sample loop) while monitoring themolecular ion m/z=314 of CBD and THC in SIM mode.

(THC) analysis by EI-LC-MS with SMB; 20 ppm each compound was injectedThe flow rate was set to 60μl/min, and the gradient was from 50% to 80%



Journal of

MASS SPECTROMETRY

As demonstrated, cleanmass chromatogram is obtained, and theVarian software determined the signal to noise ratio for THC as 737in peak to peak calculation. Thus, the linearly extrapolated LOD is0.3pg. The peak of CBD is about half of that of THC because of itslower relative abundance of its molecular ion. We note that themeasured LOD can be different from an extrapolated LOD whichis optimistic most of the times.

In Fig. 10 we show the analysis of Cannabis flower extract. A smallportion of dry Cannabis flower was crushed and extracted by puremethanol under 20-min sonication, followed by centrifugation andfiltration. The extract was injected as is into a Varian ProStar LCequipped with a 15-cm column with 1-mm ID (Phenomenex Luna,using 5-μm C18 particles) and separated with gradient elution at60μl/min solvent flow rate from 50% water/methanol mixture to100% methanol. As demonstrated, the results were initially surpris-ing as we found that the two biggest peaks had the same Cold EImass spectra that were identified by the NIST library as of THC.The explanation for this experimental observation is that the sec-ond peak is of THC-A meaning Tetrahydrocannabinol acid whichwas separated from THC on the LC column, but dissociated at theheated vaporization chamber into THC.

We analyzed four Cannabis flower samples, and as shown inTable 2 in each one of them we found the THC-A to be highly orthe most abundant compound but with THC-A/THC ratio that sig-nificantly changed from sample to sample.

This finding which is well known[56] demonstrates the impor-tance of LC-MS analysis over GC-MS (or GC-FID) and the

Table 2. The relative abundance of THC-A, THC and CBN in fourCannabis flower samples

Sample THCA THC CBN

a 79% 19% 2%

b 91% 8% 1%

c 42% 51% 7%

d 89% 9% 2%

Figure 10. Cannabis flower analysis with EI-LC-MS with SMB. The extractwas injected through a 5-μl loop to a C18 Phenomenex Luna column with5-μm particles 150mm×1.0mm by VARIAN ProStar-210 LC. The solventgradient was set to 60μl/min flow rate of 50% water for 1min andreaching 100% methanol after 5min. The vaporization chamber and thetransfer line were set to 300 °C.


effectiveness of EI-LC-MS with SMB that enables direct quantitationof the ratios of Cannabis main ingredients.

126

Discussion and advantages

A new type of electron ionization LC-MS was developed, based onthe use of supersonic molecular beams for LC and MS interface andas a medium for electron ionization of vibrationally cold samplemolecules in a fly-through ion source. The structure of this newEI-LC-MS with SMB system, and its main components is described,and its various unique benefits were demonstrated in selected ex-amples and applications. The new EI-LC-MS with SMB system pro-vides the following major benefits and advantages over currentlyused API based LC-MS:

1. Library based identification. Library searchable EI massspectra are provided, unlike with ESI and/or APCI. This is animportant advantageous feature, shared with Particle BeamLC-MS and ‘Direct EI’ that enables fast, automated, legallydefensible and reliable molecular identification with samplenames and structures at the isomer level.

2. Enhanced molecular ions. High quality Cold EI mass spec-tra are provided with enhanced molecular ions and isomerstructural information for further increased confidence levelin the sample identity, particularly versus its degradationproducts and homologous compounds.

3. Provision of elemental formula. While quadrupole massspectrometer with unit resolution is used, elemental formulainformation can be obtained in our system with the TAMIsoftware that is based on isotope abundance analysis ofthe molecular ion group of isotopomers,[48] and the similar-ity of the measured masses.

4. No ion suppression. Ion suppression and/or enhancementeffects that plague API based ion sources are fully eliminated.

5. Fast LC-MS. Faster LC-MS analysis is enabled through theavailability of automated deconvolution software and partic-ularly in view of the absence of ion suppression effects un-der coelution conditions. Consequently, full LC separationis not essential and partial co-elution can be tolerated thatfacilitates much faster analysis.

6. Flow injectionMS-MS. Fast flow injectionMS-MS can some-times replace lengthy LC-MS analysis because of the elimi-nation of ion suppression or enhancement effects.

7. Non-polar compound analysis. EI-LC-MS with SMB canserve for the analysis of both semi-polar and non-polar com-pounds and unlike ESI and/or APCI based LC-MS it can ana-lyze completely non-polar compounds including saturatedhydrocarbons and polycyclic aromatic hydrocarbons (PAHs).

8. Uniform response. Electron ionization provides uniformionization yield (response) for all compounds that enterthe ion source for improved quantitation without the needfor compounds specific calibration and it can serve for theprovision of chemical reaction yields and improved impuri-ties analysis.

9. No nitrogen generator. The bulky heavy and expensive ni-trogen generator and its air compressor (noise, service,added price and added lab or extra roomplace) that are usedin Electrospray LC-MS are not required plus the rotary pumpof the system ismuch smaller. Thus, unlike in standard LC-MSthe EI-LC-MS with SMB system can be transportable to thehood or process on a cart with a small helium cylinder.


1

B. Seemann et al.

Journal of

MASS SPECTROMETRY

1262

10. Reasonable LOD. Reasonable LOD of around or just below1pg in SIM mode were obtained for several compounds in-cluding OFN, Agidol 40, Cholesterol and Pyrene. While theseLODs are not as good as obtained with ESI-LC-MS they areabout the same for all compounds that are amenable forEI-LC-MS analysis hence better than those of ESI-LC-MS fornon-polar compounds.

11. Large linear dynamic range. High linear dynamic range(LDR) is provided of five orders of magnitude.

12. GC-MS and LC-MS with SMB in a one system. GC-MS withCold EI (SMB) and EI-LC-MS with SMB system replacementcould be obtained through a change of a one flange withthe GC-MS transfer line or LC-MS vaporization chamberand transfer line hence the system can have dual GC-MSand LC-MS usage.

13. Low cost of goods. The feature of library based identifica-tion implies that a simple low cost single quadrupole massspectrometer can be used, the same as in GC-MS, becausethe library based identification reduces the need for costlyhigh resolution MS.

Cold EI facilitates the provision of a new and useful LC-MS systemthat supplements and complements ESI-LC-MS. The EI-LC-MS withSMB excels especially in the analysis of unknown compoundsand/or for the analysis of synthetic organic compounds in whichthe uniform response enables the provision and optimization ofchemical reaction yields. The EI-LC-MS with SMB can also serve forflow injectionMS-MS to supplement and complement FI-ESI-MS-MSas a fast alternative to lengthy LC-MS analysis. Our vision is that ColdEI with its unique benefits will serve as a dedicated low cost EI-LC-MSsystem,withhighperformancefortheanalysisofsmallmolecules.It is worthwhile to briefly compare our Cold-EI with SMB ap-

proach to the Direct EI approach of Cappiello and co-workers.[20–32]

In Direct EI, the LC and MS interface is as simple as it gets, and theMS of a GC-MS can be used with only minor added EI ion-sourceand interfacemodifications. However, the LC column flow rate is re-stricted to below 1μl/min which necessitates the use of packedcapillary Nano LC columns with its resulting lower sample loadingability and loss of concentration sensitivity. We feel that the majoradvantage of our Cold EI approach is the higher quality mass spec-tra obtained which adequately compensates for the added instru-mental complexity.The EI-LC-MS with SMB system and approach require further in-

vestigation and exploration of further applications for its establish-ment as a useful technique. We also intend to combine it with othermore modern MS system such as the Agilent 5977 MSD.

Acknowledgements

This research was supported by the Israel Science Foundationfounded by the Israel Academy of Sciences and Humanities (grantno 393/11). This research was also supported by the Ministry of Sci-enceandTechnology, Israel.WethankDr.SimonProsser fromAdvionInc for the Cannabinoids samples and advise on their analysis.

References[1] M. Yamashita, J. B. Fenn. Electrospray ion source. Another variation on

the free-jet theme. J. Phys. Chem. 1984, 88, 4451–4459.[2] C. M. Whitehouse, R. N. Dreyer, M. Yamashita, J. B. Fenn. Electrospray

interface for liquid chromatographs and mass spectrometers. Anal.Chem. 1985, 57, 675–679.


[3] R. B. Cole (Ed). Electrospray Ionization Mass Spectrometry. John Wiley &Sons, Inc.: New York, 1997.

[4] D. I. Carroll, I. Dzidic, E. C. Horning, R. N. Stillwell. Atmospheric pressureionization mass spectrometry. Appl. Spectrosc. Rev. 1981, 17, 337–406.

[5] D. I. Carroll, I. Dzidic, R. N. Stillwel, M. G. Horning, E. C. Horning.Subpicogram detection system for gas phase analysis based uponatmospheric pressure ionization (API) mass spectrometry. Anal. Chem.1974, 46, 706–710.

[6] D. B. Robb, T. R. Covey, A. P. Bruins. Atmospheric pressurephotoionization: an ionization method for liquid chromatography–mass spectrometry. Anal. Chem. 2000, 72, 3653–3659.

[7] J. A. Syage, M. D. Evans. Photoionization, mass spectrometry: apowerful new tool for drug discovery. Spectrosc 2001, 16, 14–21.

[8] A. Raffaelli, A. Saba. Atmospheric pressure photoionization massspectrometry. Mass Spectrom. Rev. 2003, 22, 318–331.

[9] F. W. McLafferty. Interpretation of Mass Spectra. University ScienceBooks: Mill Valley, CA, 1980.

[10] R. C. Willoughby, R. F. Browner. Monodisperse aerosol generationinterface for combining liquid chromatography with mass spec-trometry. Anal. Chem. 1984, 56, 2625.

[11] P. C. Winkler, D. D. Perkins, W. K. Williams, R. F. Browner. Performance ofan improved monodisperse aerosol generation interface for liquidchromatography/mass spectrometry. Anal. Chem. 1988, 60, 489.

[12] R. D. Voyksner, C. S. Smith, P. C. Knox. Optimization and application ofparticle beam high performance liquid-chromatography massspectrometry to compounds of pharmaceutical interest. Biomed.Environ. Mass Spectrom. 1990, 19, 523.

[13] G. G. Jones, R. E. Pauls, R. C. Willoughby. Analysis of styrene oligomersby particle beam liquid chromatography mass spectrometry. Anal.Chem. 1991, 63, 460–463.

[14] T. D. Behymer, T. A. Bellar, W. L. Budde. Liquid-chromatography particlebeam mass spectrometry of polar compounds of environmentalinterest. Anal. Chem. 1990, 62, 1686.

[15] A. P. Tinke, R. A. M. van der Hoeven, W. M. A. Niessen, U. R. Tjaden,J. van der Greef. Some aspects of peak broadening in particle beam –

liquid chromatography – mass spectrometry. J. Chromatogr. 1991,554, 119–124.

[16] A. Cappiello, F. Bruner. Micro flow-rate particle beam interface forcapillary liquid chromatography mass-spectrometry. Anal. Chem.1993, 65, 1281.

[17] A. Cappiello, G. Famiglini. Capillary-scale particle-beam liquid chro-matography mass spectrometry interface: can electron ionizationsustain the competition? J. Am. Soc. Mass Spectrom. 1998, 9, 993–1001.

[18] A. Cappiello, M. Balogh, G. Famiglini, F. Mangani, P. Palma. An efficientliquid chromatography–mass spectrometry interface for the generationof electron ionization spectra. Anal. Chem. 2000, 72, 3841–3848.

[19] A. Cappiello, G. Famiglini, F. Mangani, P. Palma. New trends in theapplication of electron ionization to liquid chromatography–massspectrometry interfacing. Mass Spectrom. Rev. 2001, 20, 88–104.

[20] A. Cappiello, G. Famiglini, P. Palma, F. Mangani. Trace level de-termination of organophosphorus pesticides in water with the newdirect-electron ionization LC/MS interface. Anal. Chem. 2002, 74,3547–3554.

[21] A. Cappiello, G. Famiglini, F. Mangani, P. Palma. A simple approach forcoupling liquid chromatography and electron ionization mass spec-trometry. J. Am. Soc. Mass Spectrom. 2002, 13, 265–273.

[22] A. Cappiello, G. Famiglini, F. Mangani, P. Palma, A. Siviero. Nano-high-performance liquid chromatography–electron ionization mass spec-trometry approach for environmental analysis. Anal. Chim. Acta 2003,493, 125–136.

[23] A. Cappiello, G. Famiglini, P. Palma. Electron ionization for LC/MS. Anal.Chem. 2003, 75, 1173–1179.

[24] A. Cappiello, G. Famiglini, P. Palma, A. Siviero. Liquid chromatography–electron ionization mass spectrometry: fields of application andevaluation of the performance of a direct-EI interface. Mass Spectrom.Rev. 2005, 24, 978–989.

[25] G. Famiglini, P. Palma, A. Siviero, M. A. Resai, A. Cappiello. Determinationof endocrine disrupting compounds in marine water by nanoliquidchromatography/direct-electron ionization mass spectrometry. Anal.Chem. 2005, 77, 7654–7661.

[26] A. Cappiello, G. Famiglini, E. Pierini, P. Palma, H. Trufelli. Advanced liquidchromatography–mass spectrometry interface based on electronionization. Anal. Chem. 2007, 79, 5364–5372.

[27] G. Famiglini, P. Palma, E. Pierini, H. Trufelli, A. Cappiello. Organochlorinepesticides by LC-MS. Anal. Chem. 2008, 80, 3445–3449.



Journal of

MASS SPECTROMETRY

[28] A. Cappiello, G. Famiglini, P. Palma, E. Pierini, V. Termopoli, H. Trufelli.Overcoming matrix effects in liquid chromatography – massspectrometry. Anal. Chem. 2008, 80, 9343–9348.

[29] G. Famiglini, P. Palma, V. Termopoli, H. Trufelli, A. Cappiello. Single-stepLC/MS method for the simultaneous determination of GC-amenableorganochlorine and LC-amenable phenoxy acidic pesticides. Anal.Chem. 2009, 81, 7373–7378.

[30] P. Palma, G. Famiglini, H. Trufelli, A. Cappiello. Towards a universaldetector for small molecule applications: direct-El in LC-MS. LC-GCEur. 2010, 126, 23.

[31] A. Cappiello, G. Famiglini, P. Palma, H. Trufelli. Matrix effects in liquidchromatography – mass spectrometry. J. Liq. Chromatogr. RelatedTechnol. 2010, 33, 1067–1081.

[32] A. Cappiello, G. Famiglini, P. Palma, V. Termopoli, H. Trufelli. A newliquid chromatography–mass spectrometry approach for genericscreening and quantitation of potential genotoxic alkylation compoundswithout derivatization. J. Chromatogr. A 2012, 1255, 286–290.

[33] A. Amirav, A. Gordin, M. Poliak, A. B. Fialkov. Gas chromatography massspectrometry with supersonic molecular beams―special featurearticle. J. Mass Spectrom. 2008, 43, 141–163.

[34] A. B. Fialkov, U. Steiner, L. Jones, A. Amirav. A new type of GC-MS withadvanced capabilities. Int. J. Mass. Spectrom. 2006, 251, 47–58.

[35] A. Amirav, O. Granot. Liquid chromatography mass spectrometry withsupersonic molecular beams. LC–MS with electron ionization of coldmolecules in supersonic molecular beams. J. Am. Soc. Mass Spectrom.2000, 11, 587–591.

[36] O. Granot, A. Amirav. LC–MS with electron ionization of cold moleculesin supersonicmolecular beams. Int. J. Mass. Spectrom. 2005, 244, 15–28.

[37] O. Granot, A. Amirav. Electron Ionization L. C.-M. S. with SupersonicMolecular Beams. A. Cappiello (editor) in Advances in LC-MSInstrumentation. Journal of Chromatography Library. 2007, 72, Ch. 4,45–63. Elsevier, Amsterdam.

[38] A. Amirav. Mass spectrometer method and apparatus for analyzing asample in a solution. US Patent 7,247,495. 2007.

[39] A. Amirav. Capillary separated vaporization chamber and nozzle deviceand method. US Patent 8,604,424. 2013.

[40] A. Amirav, A. B. Fialkov, A. Gordin. Improved electron ionization ionsource for the detection of supersonic molecular beams. Rev. Sci.Instrum. 2002, 73, 2872–2876.

[41] B. Almagro, A. M. Ganan-Calvo, A. Canals. Preliminary characterizationand fundamental properties of aerosols generated by a flow focusingpneumatic nebulizer. J. Anal. At. Spectrom. 2004, 19, 1340–1346.

[42] A. B. Fialkov, A. Amirav. Cluster chemical ionization for improvedconfidence level in sample identification. Rapid. Com. Mass. Spectrom.2003, 17, 1326–1338.


[43] E. Kolodney, A. Amirav. Aerodynamical acceleration and rotational–vibrational temperatures in seeded supersonic molecular beams.Chem. Phys. 1983, 82, 269–283.

[44] A. Amirav, U. Even, J. Jortner. Cooling of large and heavy molecules inseeded supersonic beams. Chem. Phys. 1980, 51, 31–42.

[45] A. B. Fialkov, A. Gordin, A. Amirav. Extending the range of compoundsamenable for gas chromatography mass spectrometry analysis.J. Chromatogr. A 2003, 991, 217–240.

[46] M. Kochman, A. Gordin, P. Goldshlag, S. J. Lehotay, A. Amirav. Fast, high-sensitivity, multipesticide analysis of complexmixtures with supersonicgas chromatography–mass spectrometry. J. Chromatogr. A 2002, 974,185–212.

[47] A. Amirav, U. Keshet, T. Alon, A. B. Fialkov. By howmuch is the molecularion enhanced in Cold EI. Advanced GC-MS Blog Journal. blog.avivanalytical.com/2012/09/how-much-is-molecular-ion-enhanced-in.html, 2012.

[48] T. Alon, A. Amirav. Isotope abundance analysis method and softwarefor improved sample identification with the supersonic GC-MS. RapidCommun. Mass Spectrom. 2006, 20, 2579–2588.

[49] S. C. Nanita, A. M. Pentz, F. Q. Bramble. High-throughput pesticideresidue quantitative analysis achieved by tandem mass spectrometrywith automated flow injection. Anal. Chem. 2009, 81, 3134–3142.

[50] S. C. Nanita. High-throughput residue analysis by fast extraction anddilution flow injection mass spectrometry. Analyst 2011, 136, 285–287.

[51] S. C. Nanita, J. J. Stry, A. M. Pentz, J. P. McClory, J. H. May. Fast extractionand dilution flow injection mass spectrometry method for quantitativechemical residue screening in Food. J. Agric. Food Chem. 2011, 59,7557–7568.

[52] S. C. Nanita, N. L. T. Padivitage. Ammonium chloride salting outextraction/cleanup for trace-level quantitative analysis in food andbiological matrices by flow injection tandem mass spectrometry.Anal. Chim. Acta 2013, 768, 1–11.

[53] S. C. Nanita. Quantitativemass spectrometry independence frommatrixeffects and detector saturation achieved by flow injection analysis withreal-time infinite dilution. Anal. Chem. 2013, 85, 11866–11875.

[54] S. J. Lehotay. U.S. Department of Agriculture (USDA), AgriculturalResearch Service (ARS), 600 East Mermaid Lane, Wyndmoor, PA19038; USA (Private Communication).

[55] A. Gronewold, G. Skopp. A preliminary investigation on the distributionof cannabinoids in man. Forensic Sci. Int. 2011, 210, e7–e11.

[56] B. De Backer, B. Debrus, P. Lebrunb, L. Theunis, N. Dubois, L. Decockc,A. Verstraetec, P. Hubert, C. Charlier. Innovative development andvalidation of an HPLC/DAD method for the qualitative andquantitative determination of major cannabinoids in cannabis plantmaterial. J. Chromatogr. B 2009, 877, 4115–4124.


1263

http://avivanalytical.com/2012/09/how-much-is-molecular-ion-enhanced-in.html

http://avivanalytical.com/2012/09/how-much-is-molecular-ion-enhanced-in.html

electron ionization lc-ms with supersonic molecular beams ...polaris/papers/2015 - ei-lc-ms...

Documents