Nicolae Dincă,1Mihai -Cosmin Pascariu,2,3,4Alina Georgescu,2Loreta -Andrea Božin ,2Georgeta Simu,2Eugen Șișu2 [610277]

Nicolae Dincă,1Mihai -Cosmin Pascariu,2,3,4Alina Georgescu,2Loreta -Andrea Božin ,2Georgeta Simu,2Eugen Șișu2*
1“Aurel Vlaicu” University of Arad, Faculty of Food Engineering, Tourism and Environmental Protection, 2 Elena Drăgoi, RO -310330, Arad, ROMANIA
2“Victor Babeș” University of Medicine and Pharmacy of Timișoara, Faculty of Medicine, 2 Eftimie Murgu Sq., RO -300041, Timișoara, ROMANIA
3National Institute of Research&Development for Electrochemistry and Condensed Matter, 144 Dr. Aurel Păunescu Podeanu, RO -300569, Timișoara, ROMANIA
4“Vasile Goldiș” Western University of Arad, Faculty of Pharmacy, 86 Liviu Rebreanu, RO -310045, Arad, ROMANIA
* Correspondence: [anonimizat]
The limitations ofmass spectrometry regarding theidentification
ofisomeric structures with similar mass spectra arewell known .In
these cases, the discrimination between isomers through the
interpretation offragmentation patterns isnotfeasible, because the
mass spectra ofthese isomeric analytes contain thesame peaks .
Using spectral libraries cangive erroneous results because ofthelow
sensitivity ofthesearch algorithm inthecase ofspectra with high
similarity, orbecause oftheabsence oftheanalyte’s spectrum from
thedatabase .Wehave circumvented this problem byadditionally
using theanalytical information contained inthemain isobaric ions’
intensities .Also, thespectral library was replaced with adatabase
containing theionic orthefragmentation enthalpies corresponding to
themain isobaric ions which can begenerated from theanalytes’
possible structures .Because nounified enthalpies databases exist,
these were calculated using quantum chemical methods, usually
semi -empirical methods, which give unified values and which canbe
quickly implemented onordinary computers .
The invoked search algorithms arebased onthe kinetic and
thermodynamic fragmentation laws and use the inverse sorting
(ORD)1and theinverted linear correlation (LCOR)2oftheintensities’
row oftheisobaric ions with theenthalpies’ row.Inthis paper we
propose toestablish theoptimal ionization energy forthechemical
structure identification (CSI) ofsome acetalized sugars using the
mentioned algorithms .Weused forthis purpose five standards, two
semi -empirical methods forquantum chemical calculation (QCC) and
four ionization energies (IE).
REFERENCES
1.N.Dincă, A.Covaci .Structural identification bydifferential mass spectrometry asa
criterion forselecting thebest quantum chemical calculation offormation enthalpy for
tetrachlorinated biphenyls .Rapid Commun .Mass Spectrom .2012 ,26,2033 .
2.N.Dincă, S.Dragan, M.Dincă, E.Sisu, A.Covaci .New Quantitative Structure –
Fragmentation Relationship Strategy forChemical Structure Identification Using the
Calculated Enthalpy ofFormation asaDescriptor fortheFragments Produced inElectron
Ionization Mass Spectrometry :ACase Study with Tetrachlorinated Biphenyls .Anal.Chem .
(Washington, DC,U.S.)2014 ,86,4949 .
3.M.C.Pascariu, E.Sisu, V.L.Ordodi, L.M.Rusnac .Spectral Analysis of
Diisopropylidenated Monosaccharides .Low Energy EI-MS Fragmentation Study .Bul.
Stiint .Univ.“Politeh .”Timisoara, Ser.Chim .Ing.Mediului 2011 ,56,6.
4.HyperChem™ Professional, Hypercube, Inc.,1115 NW4thStreet, Gainesville, Florida
32601 ,USA, version 8.0.10forWindows .
5.G.B.Rocha, R.O.Freire, A.M.Simas, J.J.P.Stewart .RM1:Areparameterization ofAM1
forH,C,N,O,P,S,F,Cl,Br,andI.J.Comput .Chem .2006 ,27,1101 .
6.MOPAC 2012 ,James J.P.Stewart, Stewart Computational Chemistry, Colorado
Springs, CO,USA, http://OpenMOPAC .net/(March 2015 ),version 15.027W.
7.J.J.P.Stewart .Optimization ofparameters forsemiempirical methods VI:more
modifications totheNDDO approximations and re-optimization ofparameters .J.Mol.
Model .2013 ,19,1.
QUALITATIVE CHEMICAL ANALYSIS
POSSIBILITIES USING
DIFFERENTIAL
MASS SPECTROMETRY AND QUANTUM CHEMICAL CALCULATIONS
Figure 1.The structures and themass spectra
ofthefiveisomers .The common ions used in
CSIwere marked with reddots .Figure 2 The variation of average similarity with ionization energy
Inthesimulations made foridentification ofsome acetalized sugars, LCOR and coupling algorithms
with theRM1calculated database, placed thecorrect structures intheupper part oftheprobabilities lists.
Their accuracy isinsignificantly influenced byionization energy .This fact proves theability ofthesemi –
empirical method RM1toevaluate theheats offormation, tobecoupled with DiffMSinorder todescribe
theenergy profile oftheionization processes from mass spectrometry .
ACKNOWLEDGMENTS
This work was supported bytheRomanian National Authority forScientific Research (CNCS -UEFISCDI)
through project PN-II-PCCA -2011 -142.The accuracy oftheapplied algorithms increases slightly
with the decrease ofIE,although one would expect the
performances ofCSItodecrease since they arebased onthe
similarity ofthespectra .This surprising behavior isexplained
bythesecondary fragmentations effect “felt” bytherelative
intensities ofthethree analytical ions (m/z 245,m/z 187 and
m/z 127),which isweaker asIEdecreases .Secondary
fragmentations oftheinterest ions have asignificant share at
70eVand cause changes intheir relative intensities .This
influence isdifferent fortheanalyzed isomers, representing an
important source oferrors .AsIEdecreases, thesecondary
fragmentations oftheanalytical ions have alower share, and
theCSIerror decreases aswell.Inthecase ofisomers which
exhibit high similarities oftheir mass spectra, these errors are
insignificant also because thesecondary fragmentations are
similar and influence theintensity ofisobaric ions tothesame
extent .The preparation and characterization ofthe five reference
standards, DAF, DAG, DAGal, DAM and DAS (Figure 1),were
presented inaprevious paper .3The mass spectra were acquired with
aHPGC/MSD instrument inthepositive ionmode .The mass spectra
recorded at15eVareshown inFigure 1.
The strategy ofDfHdatabase calculus .The heats offormation
(ΔfH)ofmolecules and fragment ions database were calculated with
the semi -empirical method RM1using the HyperChem 8.0.10
software .4,5The geometries ofthe molecules and radicals were
optimized with theMM+ force field and re-optimized, using theRHF
operators formolecules orions andUHF forradicals .
TheMOPAC 2012 software6,7was used forthePM7semi -empirical
method,7with “CHARGE=+ 1”option forcations and “UHF” option for
radicals .The same HyperChem starting structures were converted to
“.ZMT” (MOPAC Z-matrix) files and run through the MOPAC 2012
interface forgeometry optimization andΔfHcalculation .
The CSI simulation was done bythe assignment offive
structures tofive spectra .Wecalculated theprobabilities oftheCSI
listusing theCSI-Diff-MS3.1.1software fortheORD algorithm .The
LCOR and thecoupling algorithms were runonordinary computers
using Microsoft Excel 2010 worksheets .CSI was conducted by
strictly adhering totheprotocols forORD, LCOR and their coupling,
presented inprevious papers,1,2forthe common isobaric ions of
compounds 1–5:[M−CH3•]+atm/z 245,[M−CH3•−acetone]+atm/z 187
and [M−HO•−2·acetone]+atm/z 127.Forthese ions theDfH(relative)
was obtained with Eqn.(1):
DfH (relative) = DfH (ion) − DfH (molecule) (1)
TheΔfH(relative) data areavailable inTables 1and2.
Table 1. Series of ΔfH (relative) values calculated using the RM1 semi –
empirical method and Eqn. (1)
Table 2. Series of ΔfH (relative) values calculated using the PM7 semi –
empirical method and Eqn. (1)
Figure 3. The variation of the CSI accuracy for ORD, LCOR and their coupling
algorithms.
Accuracy (%) = 100(N –Δ rank S ) / N (2)
where rank Sistherank ofthereal structure inthelistofprobabilities orthe
number ofstructures with theprobability greater than orequal tothat ofthe
realstructure, Δrank S=(rank S−true rank ),true rank =1;Nisthenumber of
possible structures fortheanalytes .1For5spectra and5possible structures,
N=1ˑ2ˑ3ˑ4ˑ5=120possibilities .