Application Compendium

Time-of-flight solutions in pharmaceutical development the power of accurate mass

Contents

Technique-related notes

Time-of-flight mass spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Effect of resolution and mass accuracy on empirical formula confirmation and identification of unknowns . . . . . . . . . . . . . . . . . . . . . . . . . . .14

Advantages of wide dynamic range on an orthogonal acceleration time-of-flight mass spectrometer . . . . . . . . . . . . . . . . . . . . . . . . .16

Accurate mass measurement for analyzing drugs of abuse by LC/time-of-flight mass spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20

Application-related notes

Using the Agilent LC/MSD TOF to identify unknown compounds . . . . . . . . . . .22

Automated empirical formula confirmation using the Agilent LC/MSD TOF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26

Using accurate mass LC/MS in a walk-up environment . . . . . . . . . . . . . . . . . .28

Automated empirical formula generation for the identification of unknown compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30

Screening for target compounds using the LC/MSD TOF . . . . . . . . . . . . . . . . . .32

Fast scanning for high throughput screening using the LC/MSD TOF . . . . . . .34

Structure elucidation of degradation products from the antibiotic drug amoxicillin by accurate mass measurement using LC-ESI-oaTOF MS . . . . . .36

2

Foreword

This compendium is a collection of examples and applications based on the Agilent LC/MSD time-of-flight mass spectrometer. It focuses onthe power of accurate mass in doing both qualitative and quantitativeapplications. Because this instrument makes obtaining accurate massmeasurements routine, the LC/MSD TOF has many uses in the drugdevelopment laboratory as well as other areas.

This compendium provides an overview of Agilents Application Notesand Technical Notes, currently available for using API time-of-flight-mass spectrometry. Most of the notes are presented in a condensedform, for more in-depth reading the full length versions can be down-loaded from Agilents web site at www.agilent.com/chem/tof.

Advantages of the LC/MSD TOF

Routine mass accuracy of 3 ppm or better over a broad dynamicrange to allow making accurate mass measurements without requiring special training or careful sample preparation.

Full scan sensitivity at the low picogram level enabling the identification of unexpected trace contaminants.

Linear response over three orders of magnitude dynamic range combined with high resolving power which allows for quantificationof samples in complex matrices.

A wide range of sources available including electrospray (ESI), nanoelectrospray, atmospheric pressure chemical ionization (APCI),atmospheric pressure photoionization (APPI), matrix assisted laserdesorbtion ionization (MALDI).

Sample introduction by either HPLC or capillary electrophoresis.

3

4Technique Applications

Time-of-flight mass spectrometryTime-of-flight mass spectrometry(TOF MS) was developed in the

late 1940s, but until the 1990s its

popularity was limited. Recent

improvements in TOF technology,

including orthogonal acceleration,

ion mirrors (reflectrons), and

high-speed electronics, have

significantly improved TOF

resolution. This improved resolu-

tion, combined with powerful and

easy-to-use electrospray (ESI) and

matrix-assisted laser desorption

ionization (MALDI) ion sources,

have made TOF MS a core tech-

nology for the analysis of both

small and large molecules.

This overview describes:

Basic theory of operation for

an orthogonal acceleration

time-of-flight (oa-TOF) mass

spectrometer

Flight time and the fundamental

equations for TOF mass analysis

TOF measurement cycle

Relative advantages of the two

most common TOF digitizers

analog-to-digital converter

(ADC) and time-to-digital

converter (TDC)

Theoretical and practical limits

to mass accuracy

Dynamic range considerations

Basic oa-TOF MS theory of operation

While an orthogonal accelerationtime-of-flight mass spectrometercan be interfaced with many typesof ion sources, this discussion willfocus on the use of an oa-TOF MSwith atmospheric pressure ionization(API) sources. There are severaltypes of API sources that can beused, including:

Electrospray ionization (ESI) atvarious flow rates

Atmospheric pressure chemicalionization (APCI)

Atmospheric pressure photoionization (APPI)

Atmospheric pressure matrix-assisted laser desorptionionization (AP-MALDI)

Ions from these sources can beintroduced into the mass spectro-meter vacuum system via a commonatmospheric sampling interface.

Figure 1 depicts the AgilentLC/MSD TOF, an oa-TOF massspectrometer. Ions produced inthe source are electrostaticallydrawn through heated drying gasand then through a sampling capil-lary into the first stage of the vac-uum system. Near the exit of thecapillary is a metal skimmer witha small hole. Heavier ions withgreater momentum pass throughthe skimmer aperture. Most of thelighter drying gas (nitrogen) mole-cules are pumped away by a vacu-um pump.

The ions that pass through theskimmer and enter the secondstage of the vacuum system areimmediately focused by the firstof two octopole ion guides. Anoctopole ion guide is a set of smallparallel metal rods with a com-mon open axis through which theions can pass. Radio frequency(RF) voltage applied to the rodscreates electromagnetic fields thatconfine ions above a particularmass to the open center of the rod set. The ions are propelledthrough this first octopole ionguide by the momentum retainedfrom being drawn from atmos-pheric pressure through the sam-pling capillary. As the ions transitthe first octopole, they also passinto the third stage of the vacuumsystem, where the pressure is nowlow enough that there are few collisions between the ions andgas molecules.

Ions exiting the first octopole ionguide immediately enter the secondoctopole ion guide in the fourthvacuum stage. The second octo-pole ion guide is similar to the first,but carries a lower direct current(DC) potential. It accelerates theions. The second ion guide is drivenby an RF power amplifier operatedat 5 MHz. The high 5 MHz frequen-cy is key to achieving maximum ion transmission over a wide (> m/z 100m/z 3000) mass rangeIn the fourth vacuum stage, theion beam leaves the second octo-pole ion guide and enters thebeam-shaping optics. An ion focuslens and DC quadrupole shape the

5+++

++ ++

Heated

Octopoleion

guides

Focuslens

4

Vacuum stages

Flighttube

Slits

5

DCquad

Detector

HPLCinlet

Nebulizer

Nebulizergas inlet

Analyteions

Waste

++

+

Capillary

Skimmer

Vacuumwalls

Heateddrying gas

Ion pulser

31

Signal

Ion mirror

2

Figure 1 Ion source, ion optics, and mass filter from the Agilent LC/MSD TOF, an API oa-TOF mass spectrometer.

beam to achieve optimal paral-lelism and size before it enters thetime-of-flight mass analyzer. Themore parallel the ion beam, thehigher the resolving power thatcan be achieved.

After the ions have been shapedinto a parallel beam, they passthrough a pair of slits into the fifthand last vacuum stage, where thetime-of-flight mass analysis takesplace. Because the mass of eachion is assigned based on its flighttime, the background gas pressurein this stage must be very low. Anycollision of an ion with residualbackground molecules will alterthe flight time of the ion andaffect the accuracy of its massassignment.

In the time-of-flight mass analyzer,the nearly parallel beam of ionsfirst passes into the ion pulser.The pulser is a stack of plates,each (except the back plate) witha center hole. The ions pass intothis stack from the side justbetween the back plate and thefirst plate.

To start the ions flight to thedetector, a high-voltage (HV) pulseis applied to the back plate. Thisaccelerates the ions through thestack of pulser plates.

The ions leave the ion pulser andtravel through the flight tube,which is about one meter inlength. At the opposite end of theflight tube is a two-stage, electro-static ion mirror that reversesthe direction of the ions backtowards the ion pulser. The two-stage mirror has two distinctpotential gradients, one in thebeginning section and one deeperin the mirror. This improves sec-ond-order time focusing of theions on the detector. Because ionsenter the ion pulser with a certainamount of horizontal momentum,they continue to move horizontal-ly as well as vertically during theirflight. Thus, they are not reflecteddirectly back to the ion pulsar, butinstead arrive at the detector.

6Figure 2 shows a schematic of thedetector. The first stage of thedetector is a microchannel plate(MCP), a thin plate perforated bymany precise microscopic tubes(channels). When an ion with suf-ficient energy hits the MCP, one ormore electrons are freed. Eachmicrochannel acts as an electronmultiplier. By the time the elec-trons exit the MCP, there areroughly ten electrons for everyincoming ion. The electrons exit-ing the MCP are accelerated ontoa scintillator that, when struck bythe electrons, emits photons. Thephotons from the scintillator arefocused through optical lensesonto a photomultiplier tube (PMT),which amplifies the number ofphotons and then produces a elec-trical signal proportional to thenumber of photons. The reasonfor this conversion of an electricalsignal to an optical signal andback to an electrical signal is toelectrically isolate the flight tubeand the front of the detector,which are at roughly 6,500 volts,from the PMT, whose signal out-put is at ground potential.

Flight time and its relationshipto mass

Equations for time-of-flightThe flight time for each mass isunique. It starts when a high volt-age pulse is applied to the backplate of the ion pulser and endswhen when the ion strikes thedetector. The flight time (t) isdetermined by the energy (E) towhich an ion is accelerated, thedistance (d) it has to travel, andits mass (strictly speaking itsmass-to-charge ratio).

There are two well-known formulaethat apply to time-of-flight analysis.

One is the formula for kineticenergy:

E = 1/2mv2

which solved for m looks like:

m = 2E/v2

and solved for v looks like:

v = -(2E/m)

The equation says that for a givenkinetic energy, E, smaller masseswill have larger velocities, andlarger masses will have smallervelocities. That is exactly whattakes place in the time-of-flightmass spectrometer. Ions withlower masses arrive at the detec-tor earlier, as shown in figure 3.Instead of measuring velocity, it ismuch easier to measure the time ittakes an ion to reach the detector.

The second equation is the familiarvelocity (v) equals distance (d)divide by time (t):

v = d/t

Combining the first and secondequations yields:

m = (2E/d2)t2

This gives us the basic time-of-flight relationship. For a givenenergy (E) and distance (d), themass is proportional to the squareof the flight time of the ion.

In the design of an oa-TOF massspectrometer, much effort isdevoted to holding the values ofthe energy (E) applied to the ionsand the distance (d) the ion trav-els constant, so that an accuratemeasurement of flight time willgive an accurate mass value.

Figure 2TOF detector with potentials shown for positive ion operation.

6.5KV

650V

6KV

700V

Optical lens

Scintillator

MicrochannelPlate (MCP)

ee e

e

h

Overall gain ~ 2x106

Ground

Photomultipliertube (PMT)

7As these terms are held constantthey are often combined into asingle variable, A, so:

m = At2

This is the ideal equation thatdetermines the relationshipbetween the flight time of an ionand its mass. Because the relation-ship is a squared relationship, ifthe observed flight time of the ionis doubled, the resulting mass isnot doubled, but rather it is fourtimes greater.

In practice, there is a delay fromthe time the control electronicssend a start pulse to the time thathigh voltage is present on the rearion pulser plate. There is also adelay from the time an ion reachesthe front surface of the ion detec-tor until the signal generated bythat ion is digitized by the acquisi-tion electronics. These delays arevery short, but significant.

Because the true flight time can-not be measured, it is necessary tocorrect the measured time, tm, bysubtracting the sum of both thestart and stop delay times which,when added together, are referredto as to.

t = tmto

By substitution, the basic formulathat can be applied for actual mea-surements becomes:

m = A(tmto)2

Mass calibration

To make the conversion frommeasured flight time, tm, to mass,the values of A and to must bedetermined, so a calibration is per-formed. A solution of compoundswhose masses are known withgreat accuracy is analyzed. Then,a simple table is established of the flight times and correspondingknown masses. It looks somethinglike this:

Now that m and tm are known for a number of values across themass range, the computer that is receiving data from the instru-ment does the calculations todetermine A and to. It employsnonlinear regression to find thevalues of A and to so that the rightside of the calibration equation,

m = A(tmto)2

matches as closely as possible theleft side of the equation (m), forall seven of the mass values in thecalibration mix.

Figure 3Time-of-flight analysis of ions of various masses, each with a single charge. For clarity and simplicity, this shown in a linear time-of-flight massspectrometer that does not have an ion mirror.

+

+

+

Ionpulser

Flight tube

Detector

Ionoptics

Ion source

Flight path distance (d)Accelerating

energy (E)

+

+

+

+

+

+

+

+

+ +

+

+

Table 1TOF mass calibration.

Calibrant compound Flight timemass () (sec)

118.0863 20.79841322.0481 33.53829622.029 46.12659922.0098 55.888261521.971 71.451582121.933 84.143022721.895 95.13425

8While this initial determination ofA and to is highly accurate, it isnot accurate enough to give thebest possible mass accuracy fortime-of-flight analysis. A secondcalibration step is needed. So afterthe calibration coefficients A andto have been determined, a com-parison is made between the actu-al mass values for the calibrationmasses and their calculated valuesfrom the equation. These typicallydeviate by only a few parts-per-million (ppm). Because these devi-ations are small and relativelyconstant over time, it is possibleto perform a second-pass correc-tion to achieve an even bettermass calibration. This is done withan equation that corrects the smalldeviations across the entire massrange. This correction equation, a higher-order polynomial function,is stored as part of the instrumentcalibration. The remaining masserror after this two-step calibrationmethod, neglecting all otherinstrumental factors, is typically ator below 1 ppm over the range ofcalibration masses.

Reference mass correctionAchieving an accurate mass calibration is the first step in pro-ducing accurate mass measure-ments. When the goal is to achievemass accuracies at or below the 3 ppm level, even the most minis-cule changes in energy applied tothe ions can cause a noticeablemass shift. It is possible, however,to cancel out these factors withthe use of reference mass correc-tion. With this technique, one ormore compounds of known massare introduced into the ion sourceat the same time as the samples.The instrument software constant-ly corrects the measured massesof the unknowns using the known

masses as reference. Referencemass correction is a techniquethat has been automated on theAgilent LC/MSD TOF mass spec-trometer. To introduce referencecompounds, a second nebulizerhas been integrated into the ESIion source. This reference nebuliz-er is connected to the A bottle ofthe calibrant delivery system(CDS), which is controlled viasoftware. Bottle A contains thereference compounds. The massspectrometer control software hasan editable table that contains theexact masses of these referencecompound ions. During the acqui-sition of each spectrum from thetime-of-flight analyzer, theseknown masses are identified andthe A and to values are re-opti-mized. Each stored spectrum hasits own A and to values so that thesoftware can adjust for even thesmallest instrument variation.Each spectrum is then correctedusing these values and using thecorrection equation (the higher-order polynomial function) deter-mined in the second calibrationstep described previously. Thecorrection equation needs to bedetermined only once because thesmall deviations across the massrange are nearly constant over time.

To determine the two unknowns,A and to, the reference com-pounds must contain at least twocomponents of known mass. Inorder to achieve a good fit forboth A and to, at least one refer-ence mass needs to be a low massvalue and at least one needs to bea higher mass value. For bestresults, the low m/z and high m/zreference masses should bracketthe masses of analytical interest.The reference mass correctionalgorithm for the LC/MSD TOFrequires that one mass be at or

below m/z 330 and that a secondmass be at least 500 m/z above thelow mass ion. If these conditionsare not satisfied, but at least onereference mass is found, then onlythe A term is recalculated.

TOF measurement cycle

TOF measurements do not rely onthe arrival times of ions comingfrom just a single pulse applied tothe ion pulser, but instead aresummations of the signals result-ing from many pulses. Each time ahigh voltage is applied to theplates of the ion pulser, a newspectrum called a single transientis recorded by the data acquisitionsystem. This is added to previoustransients until a predeterminednumber of sums has been made.For analyses requiring a scanspeed of one spectrum per sec-ond, approximately 10,000 tran-sients can be summed beforetransferring the data from theinstrument back to the host com-puter to be written to disk. If thetarget application involves highspeed chromatography, then fewertransients are summed, increasingthe scan speed. The mass rangelimits the number of times persecond that the ion pulser can betriggered and transients recorded.Once the ion pulser fires, it is nec-essary to wait until the last massof interest arrives at the ion detec-tor before the ion pulser is trig-gered again. Otherwise light ionstriggered from the second tran-sient could arrive before the heav-ier ions of the first transient,resulting in overlapping spectra.Table 2 shows some examplemasses with their approximateflight times and possible transientrates. These are calculated for a

9flight length of two meters and aflight potential of 6,500 volts.Under these conditions, a ion withm/z 3200 has a flight time of about0.1 milliseconds (msec), or 100microseconds (sec). Becausethere is essentially no delay timebetween transients, this meansthat 10,000 transients per secondcorrespond to a mass range of3200 m/z. For a smaller massrange, the ion pulser can be trig-gered at higher rates. For exam-ple, a mass of m/z 800 (one-fourthof 3200 m/z) reduces the flighttime to 0.1 msec/-4, or 0.05 mil-liseconds, allowing for 20,000 tran-sients per second over an 800 m/zmass range. Conversely, extendingthe transient to 0.141 millisecondsdoubles the mass range to 6400m/z (mass is a function of thetime squared).

Because transients are so short,the number of ions of a specificmass from a particular compoundin any given transient is generallyquite small. For many oa-TOFinstruments, this number averagesto substantially less than one. Thisfact plays an important role in thebasic design of the data acquisi-tion system of many of todayscommercial instruments.

Digitally recording ion arrival

While there is an exact instantwhen each ion strikes the detec-tor, it is difficult to transfer thisperfectly into the digital world.There are two basic approachesused to translate a detector signalinto a digital measurement: theanalog-to-digital converter (ADC)used in the Agilent LC/MSD TOFand the time-to-digital converterused in many other commercialTOF systems. The next two sec-tions discuss these two approaches.

Analog-to-digital converter systemsThe function of an analog-to-digi-tal converter (ADC) is to repre-sent digitally the signal that comesfrom the ion detector. An ADCdoes not attempt to determine theexact arrival time of the ions; it issimply a data recorder. As a datarecorder, it samples the amplifieddetector output at a fixed interval.In the case of the LC/MSD TOF,this interval is one nanosecond(109 seconds). This translates to afrequency of one gigahertz (GHz),or one billion cycles per second.During each cycle the detectoroutput signal intensity is convert-ed into a digital value. The digitalvalue is represented by eight bits,corresponding to a dynamic rangeof 28 counts, or in decimal nota-tion 0 to 255 counts. When theacquisition system signals thepulser to fire, the ADC begins toconvert the signal arriving fromthe detector amplifier. It storeseach successive conversion inmemory. Each time the pulserfires, the ADC adds the new mea-surement to those already record-ed in memory from the previoustransients. When an ADC is usedin this way, it is called an integrat-

ing transient recorder. With anADC, some care must be used tobias the detector amplifier (AmpOffset) to a value close to zero sothat when no ion signal is present,zero signal is recorded. Otherwise,the signal present in the absenceof an ion signal would add to sys-tem noise. The gain of the detec-tor and amplifier must be suffi-cient so that an individual ion reg-isters at least one count. In prac-tice, the gain is normally set sothat the average number of countsper ion is greater than one. TheLC/MSD TOF autotune routineautomatically sets the detectorgain and Amp Offset parametersto satisfy these conditions. Theadvantage of the ADC acquisitionsystem relative to the TDC acqui-sition system (discussed in thenext section) becomes apparentwhen multiple ions of a givenmass arrive at the detector withina single transient. The detector isan analog device and amplifies thecombined signal from the severalnearly simultaneous ion arrivals.An ADC with its eight bits cantranslate this rising and falling sig-nal into a digital profile of themass peak, as shown in figure 4.Each successive transient buildsthe values in memory. This accu-rately represents the detector out-put signal, whether it is from asmall or large ion current. Thenext section will show why theTDC does not have this dynamicrange.

Time-to-digital converter systemsThe time-to-digital converter(TDC) represents the secondapproach to digitizing a TOF sig-nal. A TDC acquisition systembegins with a discriminator. A dis-criminator is an electronic devicethat triggers when a particular sig-nal level is reached. This trigger

Table 1Flight time and transients/second as a functionof mass*.*Two-meter flight tube, flight potential 6500V.The minimum allowed transient is 50 sec(50,000 points). The maximum is 160 sec(160,000 points) or about 8,000 m/z.

m/z Flight time (sec) Transients/sec

800 50 20,0003200 100 10,0006400 141 7,070

10

signal from the discriminator isregistered by a counter, whichmarks the flight time. After a briefdead time, the discriminator andcounter are ready to record thenext ion arrival. Since the discrim-inator triggers on the leading edgeof the mass peak, the advantage ofa TDC system is its ability to elim-inate any broadening of the masspeak originating in the detectorand amplifier. One disadvantage isloss of dynamic range. Since thediscriminator triggers on the lead-ing edge of the incoming ion sig-nal, it ignores the remainder of thedetector signal and gives the sameresponse regardless of whetherthe signal is the result of one ionor many ions. The TDC simplymarks ion arrival, but cannot con-vey how many ions. Because therepetition rate for transients ishigh, and the average number ofions for any given mass has beensubstantially less than one pertransient, this has generally been an acceptable solution.However, as ion sources and ionoptics become more efficient, thenumber of ions of a given mass ina single transient increases to thepoint of significance. To illustratethis, consider a hypotheticalinstrument equivalent to theLC/MSD TOF that uses a TDCacquisition system. Figure 5 showsthe number of ions for a singlecompound that arrive in a singletransient, as a function of sampleamount. At sample concentrationsabove 1000 picograms, the hypo-thetical TDC system no longergives an increased signal responsebecause the TDC cannot reflectthe fact that multiple ions of agiven mass are arriving in eachtransient. A second problem asso-ciated with TDC acquisition sys-tems is an observed shift in mea-

sured ion arrival time at high ioncurrents. When less than one ionfor any given mass arrives at thedetector per transient, the TDCaccurately records arrival time towithin the limit of the countersresolution. If ions arrive for a givenmass just slightly separated in time(as determined by the instruments

resolving power) then, unless thesignal from the detector hasreturned to below the thresholdpoint, the second ion is unable totrigger the discriminator (see fig-ure 6). This phenomenon and theassociated reset time of the dis-criminator and counter are calledTDC dead time. TDC dead time

Mass peak

ADC sample

Ion arrival timeIn

tens

ity (#

of i

ons

reco

rded

at e

ach

arriv

al ti

me)

Figure 4An ADC can record multiple ions per transient, so it accurately tracks ion signal intensity.

0.01

0.1

Ions perTransient

Arb. Units

1E+3

1E+4

1E+5

Mean response using oa-TOF with ADC

1E+6

10 100 1000 10000Sample amount (pg)

Predicted limit for the same oa-TOF with TDC

1

Figure 5ons per transient as a function of sample amount, showing TDC limitations.

11

can have a significant effect inattempts to accurately measureaverage ion arrival times. If a sig-nificant number of ions arrive atthe detector during the TDC deadtime, then a shift in the average ofthe arrival distribution occurs. Theshift in the measured ion arrivaltime is always to shorter arrivaltimes, because it is always the sec-ond ion to arrive in a given tran-sient that is dropped. The shifttowards shorter apparent arrivaltime directly translates to a small-er mass value. When attempting tomeasure mass values to the part-per-million accuracy, even a fewions missed can have a substantialeffect. The discriminator used onTDC systems also introduces athird problem. The arrival of eachion produces a peak with measur-able width. With an ADC system,the peak is profiled with multiplepoints within a single transient.These points can be subjected tomathematical centroiding to cal-culate the arrival time with highaccuracy. Centroiding allows cal-culation of the ion arrival time toa resolution beyond that given bythe original data points. With aTDC system, the arrival of the ionis captured by a single value. Thismeans the time between dataacquisitions must be shorter toachieve the same time resolution.Because of the loss of the arrivalprofile information, TDC systemsmust operate at higher samplingrates to achieve equivalent massaccuracy even when saturationeffects are not present.

Theoretical and practical limits to mass accuracy

Whether the acquisition system isa TDC or an ADC, the arrival timefor the accumulated signal inmemory is determined by cen-troiding the mass measurementsfrom the individual transients.Even though the focus of thedesign of the TDC was to specifi-cally measure the arrival time ofeach ion, the nominal arrival timemust be the average (centroid) ofthe population for the summedtransients. There are limits to howprecisely this centroid can bedetermined.

Ion statisticsThe first theoretical limit is set bythe number of ions measured andtheir time distribution. If the dis-tribution is narrow and well popu-lated, resulting in a quiet and sta-ble signal, then the centroid oraverage can be precisely deter-mined. The expression is:

= 106/(2.4 * R * -n)

where is the standard deviationof the resulting measurement, R isthe resolving power (often calledresolution) of the mass spectrom-eter and n is the number of ionsthat are detected in the masspeak. Suppose one desires 95 %confidence 2 mass accuracy at 3 ppm. Then with a resolving powerof 10,000 (and 1 =1.5 ppm) it isnecessary to have approximately1000 ions. To increase the numberof ions in a centroided spectrum,it is general practice to use thedata analysis software to averagespectra across the width of theeluting chromatographic peak. Itshould be noted that while oa-TOFhas the potential for fast scancycles, reducing the scan timereduces the number of transients,which reduces the integration ofions required to achieve accuratemass measurements. Fast scan-ning and accurate mass are oppos-ing performance goals. The mostaccurate mass measurements areachieved under slower scanningconditions.

TDCthreshold

Ion arrival time:50.000000 sec

First ion arrivesat detector and is recorded.

Second ion of same nominal mass arrivesat detector during TDCdead time and is not recorded.

Inte

nsity

(# o

f ion

s re

cord

ed a

t eac

h ar

rival

tim

e)

Mass peak

Figure 6TDC dead time causes shift to shorter arrival times for higher signal levels.

12

Chemical backgroundThe second significant factor thatlimits mass accuracy is chemicalbackground. The high resolvingpower of a TOF system helps toreduce the chances of having thepeak of interest merged withbackground, yet even a smallunresolved impurity can shift thecentroid of the expected mass.The magnitude of this effect can beestimated by using a simpleweighted average calculation.

obs = contaminant1/2Abdcontaminant

(Abdcontaminant + Abdsample)

where

obs is the observed shift in massin ppm

contaminant is the mass differencebetween the sample and contami-nant in ppm

Abdcontaminant and Abdsample arethe mass peak heights or areas ofthe contaminant and sample

By way of example, for a resolvingpower of 10,000, a mass differencebetween the sample and contami-nant of 50 ppm, and relative masspeak heights of 10:1 (sample vsbackground) the observed massshift would be 50 1/(1 + 10) orabout 5 ppm. There are a numberof ways to minimize chemicalbackground. First, the AgilentLC/MSD TOF has a sealed ionsource design that minimizes cont-amination from the laboratory air.Second, very high purity HPLCsolvents should always be used.Third, a regular, systematic clean-ing program for the HPLC and theMS ion source should be followed.

These precautions help ensure thehighest quality mass measure-ments.

Dynamic range

Dynamic range can be measuredin various ways. Probably themost exacting definition for massspectrometry is the in-scan con-dition. This is the dynamic rangewithin a single spectrum, definedas the ratio in signal abundance ofthe largest and smallest usefulmass peaks.

Even when restricted to the in-scan definition of dynamic range,the upper and lower limits mustbe defined. There are both theo-retical and practical limits to con-sider. Theoretically, it is possibleto detect a single ion, but practi-cally, chemical background would,under most conditions, obscuresuch a low level. Practical limita-tions depend on the application.For example, when the instrumentis used for accurate mass mea-surement, then the lower limit isset by the minimum sampleamount for which accurate massmeasurements can be obtained. Todetermine the minimum sampleamount, the limitations based onion statistics must be considered.Assuming a goal of 5 ppm massaccuracy, achieved with 67 % con-fidence (1) based on a singleunaveraged spectrum and allow-ing for 1 ppm of calibration error,then 1 = 4 ppm. Staying with theassumption of 10,000 resolvingpower, then about 200 ions arerequired for the measurement.This calculation is based on ion

statistics and resolving power, andis independent of acquisition tech-nology. This calculation doesassume that there is significantsensitivity (signal-to-noise) so thatthe measurement is unaffected bybackground contamination. Todetermine the highest level underwhich accurate mass measure-ments can be obtained, the type ofacquisition system must be con-sidered. With a TDC system, thereis a theoretical limit at one ion pertransient at a given mass. With anADC system, depending on thedetector gain, many ions can beaccurately measured for a givenmass in a single transient. TheLC/MSD TOF autotune softwaretargets the detector gain for amean ion response of five counts.In a single transient, the ADC with8 bits or 255 counts can thereforemeasure up to 50 ions for a givenmass. Practical considerationslimit both TDC and ADC systemswith regards to the upper limit forwhich accurate mass measure-ments can be achieved. For a TDCsystem, long before the level ofone ion for a given mass per tran-sient is reached, substantial massshifts are observed. Deadtime cor-rection algorithms compensate forthis, but these corrections areeffective only up to some fractionof this theoretical limit, typically0.2 to 0.5 ions/transient. Both ADCand TDC systems, when used tomake measurements on rising andfalling chromatographic peaks,need to allow for a safety buffer ofa factor of two. This is becausethe chromatographic peak may berising into saturation, even whilethe average of the 10,000 tran-sients used to make the final massmeasurement is at only the 50 %

13

level. Table 3 summarizes boththeoretical and practical dynamicrange limits for ADC- and TDC-based oa-TOF mass spectrome-ters, based on single- spectrum, in-scan dynamic range. Dependingon the application, it is sometimespossible to extend the practicaldynamic range. One approach is tosum (average) multiple spectratogether. This improves ion statis-tics and allows for increased massaccuracy at lower sample levels. Toextend the dynamic range on thehigh end, the opposite approach istaken and spectra from the apexof a chromatographic peak areexcluded from the average.Intelligent spectral averaging is animportant function of the automat-ed accurate mass report genera-tion software of the LC/MSD TOF.Together these techniques canextend the practical limit ofdynamic range (~103) by a factorof 100, achieving effective dynamicranges of 105 for ADC-based systemsin accurate mass applications.

Conclusion

Over the past few years, there hasbeen substantial progress in tech-nologies that take the oa-TOF tonew performance levels. High-effi-ciency ion optics and vacuum sys-tem designs have given rise togreater sensitivities. High-speedADC-based acquisition systemshave made greater mass accuracyand wider dynamic range possible.The addition of sophisticated datasystems and data processing algo-rithms has enabled outstandingmass accuracies under routineanalysis conditions. By under-standing the concepts of oa-TOFmass spectrometry, it is possibleto achieve the ultimate in perfor-mance with the Agilent LC/MSDTOF system.

LC/MSD HypotheticalTOF TDC system

Theoretical limitMinimum detectable per spectrum (ions/spectrum) 1 1Maximum detectable per transient (ions/transient) 50 1Maximum detectable per spectrum ( 10,000 transients) 500,000 10,000Dynamic range 500,000 10,000

Practical limit (while achieving accurate mass)Lower limit per spectrum (ions/spectrum) 200 200Upper limit per transient (ions/transient) 25 0.10.25Upper limit per spectrum ( 10,000 transients) 250,000 10002500Dynamic range 1,250 1025

Table 3Single spectrum in-scan dynamic range.

14

Technique Applications

Abun

danc

e

Low resolving power

Target mass

Interference

Mass

Abun

danc

e

High resolving power

Target mass

Interference

Figure 1The high resolving power of a TOF mass analyzer helps to reduce the chances of having themass peak of interest merged with an interfering ion from the sample or the background.

Effect of resolution and mass accuracy on empirical formula confirmation and identification of unknowns

Mass resolution and mass accura-

cy are both critical aspects of MS

performance. With sufficient mass

resolution and mass accuracy, a

mass spectrometer can positively

confirm elemental composition or

identify unknowns.

The Agilent LC/MSD TOF design

includes unique design features

that enhance both its mass resolu-

tion and mass accuracy.

Design elements that enhanceresolving power include: A beam shaper and related ion

optics that reduce variations inion position and energy beforethey enter the mass analyzer.

One-dimensional harp gridsoriented in the direction of iontravel in both the pulser and ionmirror (reflectron).

A mechanical design that auto-matically creates proper align-ment (parallelism).

Design elements for mass accuracyinclude: An analog-to digital (ADC)

acquisition system that providesseveral orders of magnitude ofdynamic range.

15

Detailed note: Effect of resolution and mass accuracy on empirical formula confirmation and identification ofunknowns, Agilent Technologies Technical Overview, publication number 5989-1052EN (2004).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

An automated calibrant deliverysystem and a second nebulizerthat allow the continuous intro-duction of a reference masscompound at a concentrationlow enough that it is unlikely tointerfere with analyses.

A flight tube made from a spe-cial, ultralow-thermal-expansionalloy that minimizes flight pathchanges due to temperaturechanges.

Mechanical and electronic tem-perature compensation in theflight tube and electronics.

Even if mass resolution and massaccuracy are not sufficient forpositive identification, accuratemass measurements can reducethe number of likely candidatesenough that positive identificationcan be made based on a combina-tion of accurate mass measure-ments and other information.Information that may help to limitsome of the elemental composi-tion possibilities to those thatmake chemical sense can includestarting materials, the isotope dis-tribution, the number of possiblenitrogens and the number ofunsaturated bonds in a compound.In the case of a target compound,the expected empirical formula isknown and can be comparedagainst the measured accuratemass data to confirm identity.

Abun

danc

e Low resolving power

High resolving power

Mass

MeasuredMass

CalculatedMass Mass

Error

Figure 2Mass accuracy is independent of resolving power, but the LC/MSD TOF exhibits outstandingresolving power and mass accuracy.

Advantages of wide dynamic range on an orthogonal accelerationtime-of-flight mass spectrometer

16

Time-of-flight mass spectrometry

A wide dynamic range is a desir-

able characteristic in mass spec-

trometers. It allows compounds of

differing abundances to be ana-

lyzed at the same time. This can

simplify sample preparation and

facilitates walk-up access for

researchers who are not experts

in mass spectrometry. In the case

of time-of-flight mass spectrome-

ters, it also facilitates the intro-

duction of a reference mass com-

pound, which enhances the accu-

racy of mass measurements.

This technical overview reviews

the concept of dynamic range. It

then discusses how dynamic range

is achieved in a time-of-flight mass

spectrometer and how a detector

using analog-to-digital converter

technology can dramatically

improve dynamic range compared

to older detector technologies.

Dynamic range

The ratio of the maximum signal,which is the upper limit, to theminimum signal, which is thelower limit, to the magnitude dis-tinguished from background, isthe mass spectrometer's dynamicrange.

Effect of dynamic range on MS

Dynamic range has a significanteffect on a mass spectrometersusability. If an instrumentsdynamic range is very narrow it iseasy to introduce the wrongamount of sample and either get aclipped signal (too much sample)or no signal (too little sample). If mixtures of chemicals are being

analyzed, a narrow dynamic rangemeans that all of the compoundsin the mixture must have similarabundances. If they do not, eitherthe signals from high-abundancecompounds are clipped, affectingresponse linearity and makingquantitation difficult, or signalsfrom low-abundance compoundsare not detected and those com-pounds are not identified (figure 1).

In contrast, a wide dynamic rangemakes it much easier to use amass spectrometer (figure 2).Sample preparation and abun-dance becomes less critical. Ifmixtures are being analyzed, bothhigh- and low abundance com-pounds can be identified. Linearityis improved, making quantitationeasier and more reliable.

Technique Applications

Figure 1Consequences of narrow dynamic range.

17

Implications of dynamic rangefor TOF MS

For time-of-flight (TOF) massspectrometers, a wide dynamicrange has an additional benefit. Tomaximize mass accuracy, a refer-ence mass compound whose massis known to a very accuratedegree is often introduced into aTOF MS for purposes of calibra-tion. If a TOF MS has a narrowdynamic range, the reference masscompound must be introduced atan abundance nearly the same asthe abundance of samples beinganalyzed. This can create signifi-cant chemical interferences andaffect analytical results.

If, on the other hand, a TOF MShas a wide dynamic range, a refer-ence mass compound can beintroduced at an abundance muchlower than the abundances of typi-cal samples. This eliminates, or atleast minimizes, interferences.Thus, a reference mass compoundcan be introduced continuouslyand the mass accuracy of the TOFMS never has a chance to drift,ensuring maximum mass accuracy(figure 3).

TOF MS design and dynamicrange

The dynamic range of a TOF MS isdetermined by many aspects ofthe instrument design. One of themost influential is the detectorand its electronics. The detectorsenses ions as they impact it andgenerates a signal based on thoseimpacts. The signal is converted todigital form so that it can beprocessed. A time-of-flight mass

Figure 2Advantages of wide dynamic range.

250

Arbi

trary

Uni

ts

400x

1.6 ppmmass error

Sample Peakm/z 582.319

Reference Massm/z 922.010

100,

000

Arbi

trary

Uni

ts

Figure 3Wide dynamic range enables continuous introduction of reference mass compound at a lowlevel, providing maximum mass accuracy with minimum interference.

18

spectrometer determines ionmasses based on the time it takesthem to fly from a starting pointto an ending point. TOF is apulsed technique, with each elec-trical pulse starting a group ofions on their flight to the detector.For each pulse, the detectorrecords a corresponding spec-trum, called a transient. Manytransients are summed to create amass spectrum.

Time-to-digital signal conversionTraditionally, TOF mass spectrom-eters have used a detector with atime-to-digital converter (TDC). ATDC records a precise arrival timefor each mass within a transient.Because a TDC records the arrivaltime based on the arrival of thefirst ion of a given mass, it cannottell the difference between oneion at a given mass arriving in atransient and several ions with thesame mass arriving in a transient(see figure 4). This was acceptablein previous generations of TOFmass spectrometers, when ineffi-cient ion sources and ion opticsmade it unlikely that more thanone ion of a given mass would bepresent in a single transient.However, with the improved effi-ciency of modern ion sources andion optics, there is a good chancethat a transient will include multi-ple ions at a given mass. TheTDCs inability to determine howmany ions are arriving severelylimits the instruments dynamicrange. TOF mass spectrometersusing TDC detectors often havedynamic ranges of only one or twoorders of magnitude. Sampleabundance must be adjusted to

match the instruments range.Often this entails a time-consum-ing trial-and-error, dilute-and-reshoot process. One partial solu-tion to TDC dynamic range limita-tions is beam-splitting or defocus-ing. While this helps preventdetector saturation, it also resultsin a large portion of the ions beingdiscarded. This can reduce sensi-tivity and cause low-abundancecomponents of a sample to be lostentirely.

Analog-to-digital signal conversionIn contrast, the Agilent LC/MSDTOF mass spectrometer uses a dif-ferent approach, a detector withan analog-to-digital converter(ADC). The ADC does not try torecord arrival times for individualions. Instead, it records the totalsignal strength the detector is out-putting at fixed, and very frequent,intervals as often as a billiontimes per second. The ADCrecords differing signal levels, soit can tell the difference betweenone ion of a given mass or manyions of a given mass arriving in a

Sulfadimethoxine (C12H14N4O4S) and Sulfamethazine (C12H14N4O2S)with a 40 fold difference in concentration

150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350m/z

1.0e4

1.0e5

2.0e5

Inte

nsity

, cou

nts

311.08102

279.09107

0.470 ppm error

260 270 280 2900

1000

2000

3000

4000

5000

Inte

nsity

, cou

nts

279.09107

0.2718ppm error

Figure 4Wide dynamic range makes it easier to detect lower-abundance compounds in the presence of

higher-abundance compounds.

19

single transient. The result of thisis that the LC/MSD TOF, with itsADC detector, has a dynamicrange of three to four orders ofmagnitude. A resulting advantagecompared to older TOF MSdesigns is that the LC/MSD TOFfeatures continuous introductionof reference mass compound at avery low abundance. This greatlyenhances the mass accuracy ofthe LC/MSD TOF (better than 3 ppm) while minimizing interfer-ence with actual samples. Thewide dynamic range also enablesthe LC/MSD to detect less-abun-dant compounds in the presenceof significantly more abundantcompounds, even within a singlescan (see figure 4). This can be ofgreat benefit when trying to identi-fy minor impurities in the prod-ucts of synthetic chemistry andpurification, or when looking forpost-translationally modified pro-teins in the presence of native pro-teins.

Conclusion

A wide dynamic range providessignificant benefits for mass spec-trometers, facilitating the detec-tion of lower-abundance com-pounds in the presence of higher-abundance compounds. For time-of-flight mass spectrometers, thisallows the continuous introduc-tion of reference mass compoundsthat greatly enhance mass accuracy.The Agilent LC/MSD TOF featuresa detector with analog-to-digitalsignal conversion that greatlyenhances its dynamic ranges compared to older TOF designsthat rely on time-to-digital signalconversion.

Detailed note: Advantages of wide dynamic range on an orthogonal acceleration time-of-flight mass spectrometer,Agilent Technologies Technical Overview, publication number 5989-1728EN (2004).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

20

Technique Applications

1.0 9.0 12.0 13.011.010.08.07.06.05.04.03.02.00.0Time (min)

2.0e63.0e6

4.0e65.0e66.0e66.0e67.0e68.0e6

1,

68

7

910

11

14

13

Inte

nsity

(cps

) 2,3

4 5 12

0.0

Figure 1TIC of basic DA. LC/ES-MS TOF separation of 14 basic DA used as targeting compounds in toxicological sample screening.

3.373.81

1.0 9.0 12.0 13.011.010.08.07.06.05.04.03.02.00Time (min)

2.0e6

3.0e6

4.0e65.0e66.0e6

1.0e60.0

7.0e6

Inte

nsity

(cps

) Hydrocodone

Codeine

Figure 2Extracted ion chromatogram. Ion chromatogram of m/z 300.0-300.2 extracted from the data shownin figure 1. Isomers are chromatographically separated, facilitating their identification.

Accurate mass measurement for analyzing drugs of abuse by LC/time-of-flight mass spectrometry

The Agilent LC/MSD TOF provides

routine and seamless accurate

mass measurement for unambiguous

identification of chemical sub-

stances of forensic relevance like

drugs of abuse.

14 basic drugs of abuse were sepa-rated using the LC/MSD TOF.Although in some cases chromato-graphic separation is not complete(figure 1), examination of extract-ed ion profiles for each of the[M+H]+ ions shows that eachcompound can be identified with-out interference as in the case ofthe isomers, hydrocodone andcodeine (figure 2).

Table 1 shows the mass accuracyachieved from 50 pg of these com-pounds to 50-ng injected on-col-umn. Accuracy of better than 5ppm is achieved for amphetamineand better than 2 ppm for oxy-codone. Note that for low massmeasurements, the number of pos-sible empirical formulas is far lessand a 5-ppm range is more thansufficient. At a higher mass, thepossibilities increase and a lower

Peak Number 3 5 1 4 2

Compound Amphetamine Methanphetamine Hydrocodone Codeine OxycodoneNominal (m/z) 136.10 150.10 300.15 300.15 316.15

Conc. Measured Measured Measured Measured Measured(ng-injected) error (ppm) error (ppm) error (ppm) error (ppm) error (ppm)

50.00 4.97 2.41 0.94 0.94 1.3225.00 3.53 2.49 1.17 1.60 0.27

5.00 4.71 3.01 0.37 0.37 0.045.00 4.53 3.03 0.47 0.37 0.165.00 4.53 3.05 0.30 0.30 0.752.50 5.00 2.78 1.60 1.60 0.420.50 4.23 2.20 1.34 1.90 0.040.50 5.01 2.48 1.11 1.27 0.330.25 5.70 2.69 0.95 1.27 1.160.05 5.00 5.42 3.60 2.26 0.53

Table 1Accurate mass measurements vs. concentration of some DA in reference material using targeted automatic search of empirical formula.

21

range for error is needed to pro-vide confirmation or suggest a rea-sonable empirical formula to aidthe identification of an unknown.

Figure 3 shows a response vs. con-centration plot for codeine. Thehighly-linear response indicatesthis instrument can also be usedfor quantitative analysis. Note thatthe 50-ng injection was excluded,with this compound and others,because of detector saturation.For these compounds at saturatedconcentrations, accurate massmeasurement was made at theedges of the chromatographicpeak with an automated script.Also, a detection limit was not setand the 50-pg injection was madeas an arbitrary low standard. TheLC/MSD TOF specification forreserpine is 10 pg at a signal tonoise ratio of 10:1. With the instru-ments high mass resolution andseamless auto-calibration of everyspectrum collected, selectivity ofthe extracted ion is increased.

Visit our website at www.agilent.com/chem/tof to download the detailed note.

Detailed note: Accurate mass measurement for analyzing drugs of abuse by LC/time-of-flight mass spectrometry,Agilent Technologies Technical Overview, publication number 5989-0667EN (2004).

300000

200000

100000

0

0 5 10Concentration (ng/L)

15 20 25Ar

ea

Figure 3Codeine linearity. Plot of codeine extracted ion (ms/300.0-300.2) chromatographic peaks measured from 50-pg injected on-column to 25-ng on-column. TOF detector saturated at 50 ng.

Conclusion

Very high sensitivity is achievedand, with the TOF detection, alldata are full scan, allowing com-pounds that are not targeted to bedetected. The system offers a wide dynamic range capable ofproviding accurate mass measure-ments across that range withouthaving to match lock-mass signalintensity with analyte intensity.Finally, a linear response isachieved within a concentrationrange below detector and electro-spray saturation.

22

In the pharmaceutical industry, it

is not unusual to generate small

quantities of unknown compounds

in addition to the intended prod-

ucts when synthesizing lead mole-

cules, when doing larger scale

synthesis of promising candidates,

or during manufacturing. While

HPLC with a UV detector can

detect the presence of unknowns,

mass spectrometry is usually

required to positively identify

them. Since the synthesis or

manufacturing process generally

provides some clues to the com-

position of unknowns, it is fre-

quently possible to propose and/or

confirm a logical structure. One

approach to identifying these

unknowns is to interpret the com-

bination of precursor ion spectra

and product ion spectra produced

by MS/MS analysis. An alternate

approach is to use the accurate

mass capabilities of a time-of-

flight (TOF) mass spectrometer.

Accurate mass TOF systems are

used to confirm target com-

pounds,1 but can also be used to

identify unknowns.

While a quadrupole mass spec-trometer is typically operated withunit mass resolution and canassign the mass to the nearest 0.1 m/z, modern TOF systems pro-duce spectra with a resolution of4000 to 10,000, depending on themass of the ion, and can assign amass to better than 5 ppm accuracy.This has significant implicationswhen trying to propose possibleempirical formulas. Table 1 showswhat happens when you considera molecule such as reserpine (MW608.2734) and restrict the elemen-tal composition to combinationsof C, H, O, and N.

Note that even with 2 ppm massmeasurement accuracy, there arestill two possible molecular for-mulas. Furthermore, a unique formula does not translate to aunique structure. Typically, otherinformation such as knowledge ofthe synthesis, other spectral data,or product ion spectra from colli-sion-induced dissociation (CID) isneeded for unambiguous identifi-cation.

Experimental

The system used was an AgilentLC/MSD TOF coupled to anAgilent 110 Series HPLC.For more information about theapplied experimental conditionssee the original Application Note.

BPCs and EICs covering a massrange narrower than the completescan range often gave signals thatmore clearly showed the presenceof trace compounds. Figure 1shows an example of a TIC (a)and an EIC (b) from the samedata. The vertical scale on the EICis also expanded 10 times. Spectrafrom the appropriate chro-matogram were averaged and sus-pected molecular ions were select-ed. Using the elemental composi-tion calculator built into the dataanalysis software with a toleranceaccuracy set to 3 ppm, and apply-ing constraints on the elementalcomposition, possible formulaswere determined for simulatedunknowns, where the actual com-pound was known, and for trueunknowns that were trace compo-nents in synthetic mixtures.

Results and discussion

The first sample analyzed was theantibiotic, chloramphenicol, whichwas treated as an unknown inorder to demonstrate the generaltechnique. The elements wererestricted to C, H, O, N, S and Cl

Using the Agilent LC/MSD TOF to identify unknown compounds

Table 1Number of theoretical formulas for a compoundcomposed of C, H, O, and N with a molecularweight 608.2734.

Mass Accuracy (ppm) Possible Formulas

165 (quadrupole) 20910 135 73 42 2

Technique Applications

23

and the number of charges was 1because the compound was ana-lyzed in negative ion mode.

The elemental composition calcu-lator produced four possible for-mulas based on the ion at m/z321.0046 as shown in figure 2. Inorder to narrow the choices, theisotope ion at m/z 322.0081 shouldbe consistent with the number ofcarbons in the molecule and thenaturally occurring abundance of

the 13C isotope. In addition, theion at m/z 323.0017 should reflectthe naturally occurring abun-dances of 34S and 37Cl. SelectingShow Isotopic from the calcula-tor will actually overlay the theo-retical isotope abundances on thespectrum. Taking these criteriainto account, the first formula,C11H11N2O5Cl2 turned out to bethe correct formula.

Figure 1TIC and EIC of a mixture with lower-level unknown components.

24

In addition to isotope information,adduct information can be used toreduce the number of possible for-mulas. For example, a mixture offour sulfa drugs: sulfamethizole,sulfamethazine, sulfachloropyri-dazine, and sulfadimethoxine,gave a chromatographic peak thatproduced a mass spectrum withions at m/z 311.0814 and 333.0624as shown in Table 2. Using the ele-mental composition calculator thefollowing possible formulas wereproposed, assuming the m/z333.0624 ion could contain sodium.

Choice A has too few carbons tomatch the isotope data. Choices Band E are the only formulas thatare consistent with a protonatedand sodiated adduct.

Even in instruments that are notcapable of MS/MS, in-source CIDcan yield structural informationthat can be used to aid in the iden-tification of unknowns. An analysisof the sulfonamide antibiotic, sul-fachloropyridazine, demonstratesthe usefulness of this approach.The LC/MSD TOF can acquirealternating scans at different frag-mentor voltages. By using valuesof 150 and 215 volts, spectra withand without fragmentation wereobtained across the chromato-graphic peak representing sul-

Figure 2Possible formulas generated from chloramphenicol data by the elemental composition calculator.

Table 2Possible ion formulas consistent with observedions.

Ion at m/z 311.0814 Ion at m/z 333.0624

A C5H15N10O2S2 D C10H9N10O2S B C12H15N4O4S E C12H14N4O4SNa C C13H11N8S

25

fachloropyridazine. Figure 3 showsa resulting mass spectrum. In addi-tion to the protonated and sodiat-ed ions at m/z 285.0207 and307.0027, there is a fragment ion atm/z 156.0016. By observing the iso-tope ratios, it is possible to narrowdown the number of proposed for-mulas from seven to one. Since theion at m/z 156.0016 must comefrom the parent ion, its list of pos-sible elements is based on the fiveelements (C, H, N, O, S) that makeup sulfachloropyridazine. With thisconstraint, the calculator onlycomes up with a single formula,[C6H6NO2S]

+ to match the data.

Conclusion

The identification of unknowncomponents in synthetic mixturesis made much easier by theAgilent LC/MSD TOF. Mass accu-racy of better than 3 ppm resultsin a fewer potential formulas formost compounds. When otherinformation such as isotope ratios,fragment ions, and adduct ion are considered, it is often possibleto propose a unique empirical formula.

Detailed note: Using the Agilent LC/MSD TOF to identify unknown compounds, Agilent Technologies ApplicationNote, publication number 5989-0626EN (2004).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

+TOF MS: Experiment 2, 0.932 to 1.007 min from sulfa 284 a.wiff Agilent

Max. 4.8e4 counts

150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320m/z (amu)

0%5%

10%15%20%25%30%35%40%45%50%55%60%65%70%75%80%85%90%95%

100%

307.0027

285.0207

149.0242 309.0000

Rel.

Int.

(%)

156.0116

Figure 3Sulfachloropyridazine spectrum with adduct and fragment ions.

Automated empirical formula confirmation using the Agilent LC/MSD TOF

26

time-of-flight (TOF) mass spectro-

meters. The Agilent LC/MSD TOF

mass spectrometer has features

and attributes that make it much

closer to quadrupole mass spec-

trometers in its ability to work

more reliably over a wide range

of sample conditions. This note

describes the use of the Agilent

LC/MSD TOF for automated empiri-

cal formula confirmation in a high-

throughput environment.

Experimental

A detailed technical overview ofthe LC/MSD TOF is available inthis compendium on pages 4-13. A high-throughput HPLC systemwith alternating column regenera-tion was used. The sample usedfor this experiment contained sulfamethizole, sulfamethazine,sulfachloropridazine and sul-fadimethoxine in a 96-well plate atthe 10 ng/g level. The experimentalconditions are described in detailin the full length Application Note.

Results and discussion

Empirical formula confirmationThe data processing method auto-matically calculated the targetmolecular weight from the formula.An extracted ion chromatogram(EIC) was produced to cover the

mass range of all chosen adducts.An average mass spectrum wastaken across the largest peak inthe EIC and the masses for eachtarget were compared against thetheoretical mass. A one-pagereport was produced for each tar-get compound showing the EIC,the average mass spectrum, anexpanded mass spectrum showingthe adduct ions, and the resultsfor each adduct in both absolutemass error and ppm error. Anexample of this report is shown infigure 1. Early results showed thesystem typically provided accuratemass to within 3 ppm for sampleamounts in the low picogramrange to low nanogram range. Atthe high end, the detector eventu-ally saturated resulting in distort-ed mass peaks that were difficultto assign mass and yielded poorisotope ratios. The data process-ing method was modified to evalu-ate the spectra across the EICpeak for saturation. The user canset a threshold (50% was used inthese experiments) and only spec-tra below this threshold are con-sidered in producing the averagespectrum used for the subsequentcalculations. This significantlyincreased the dynamic range overwhich samples could be run with-out diluting and reinjecting. Thedata from the four 96-well plateswere processed in Excel. Themass error was plotted against signal abundance and molecularweight. The average error was lessthan 3 ppm RMS for each plate.Some plates gave less than 2 ppm

Technique Applications

In the process of drug discovery, a

major objective is the generation of

lead molecules. Laboratories may

synthesize anywhere from a few

new molecules per week to thou-

sands of molecules per week.

Typically, before these lead mole-

cules are screened for suitability,

mass spectrometry is employed to

confirm that the intended mole-

cules were synthesized, because

screening an incorrect compound

can waste time and money. Most of

this confirmation has been done by

LC/MS, using single quadrupole

instruments that have only unit

mass resolution. The problem with

this approach is that while negative

results prove the compound was

not synthesized, positive results at

this resolution only indicate a high

probability that the correct com-

pound was synthesized. If, instead,

a mass spectrometer capable of

mass accuracy better than 5 ppm is

used, the chemist will have a much

higher confidence that the correct

compound was made when a posi-

tive result is obtained.

Quadrupole mass spectrometers

are generally accepted as routine

tools, and have been refined

to work over a broad range of sam-

ple conditions with little user inter-

vention. Historically, the same

has not been true of high-resolu-

tion mass spectrometers such as

27

error. For one plate, the error onsamples below molecular weight250 was between 4 and 10 ppm.However, when the possible for-mulas were calculated for themeasured mass, each sample gavea unique formula within the calcu-lated mass error.

High-throughput resultsThe injection volume of the sam-ple described was 0.5 l. This pro-duced four separate peaks, eachapproximately three secondswide, eluting in less than 0.8 min-utes (see figure 2). The systemacquired 15 scans across eachpeak. Based on 12 runs, the massaccuracy was 0.8 ppm RMS.By using automatic column regen-eration, as well as the overlappedinjection and automatic delay-vol-ume reduction features of the wellplate sampler, the cycle time perrun was one minute and 23 sec-onds including the time requiredto process and produce the empir-ical formula confirmation report.

Conclusion

The Agilent LC/MSD TOF is capa-ble of reliably performing auto-mated empirical formula confir-mation over a broad range of sam-ple concentrations and molecularweights. Mass accuracy averagedbetter than 3 ppm. Using short,small-particle-size columns at highflow with automatic columnregeneration, the system had suffi-cient scan speed to confirm theempirical formula of a 96-wellplate library in just over 2 hours.

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Time (min)

0.02.0e44.0e46.0e48.0e41.0e51.2e51.4e51.6e51.8e52.0e52.2e52.4e52.6e52.8e53.0e53.2e53.4e53.6e5

Inte

nsity

, cps

B

CA

D

Column 1 Column 2

Empirical Formula: C10H9ClN4O2S Exact Mass: 284.01347 Sample Name: 10 ng sulfa Sample ID: Data File Name: p:\Projects\sulfa 2\Data\10ng_sulfa.wiff Acq Time: May 08 2003, 12:05:38 PM

Species Abundance (counts) Target Mass (amu) Measured Mass (amu) Mass Error (m/z) Mass Error (ppm)

[M+H]+ 59329.42 285.02075 285.02019 .00056 1.97

[M+Na]+ 2799.10 307.00270 307.00248 .00021 .70

Page of 1Empirical Formula Confirmation Report

Thursday, May 8, 2003Page 1 of 1

XIC of +TOF MS: 284.9 to 307.1 amu from 10ng_sulfa.wiff Max. 7.0e5 cps.

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5Time, min

0.0

1.0e5

2.0e5

3.0e5

4.0e5

5.0e5

6.0e5

7.0e5

Inte

nsity

, cps

285.02

157.03

+TOF MS: 3.773 to 4.124 min from 10ng_sulfa.wiff Agilent Max. 5.9e4 counts.

150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950m/z amu

0.0

1.0e4

2.0e4

3.0e4

4.0e4

5.0e4

5.9e4

Inte

nsity

, cou

nts

285.0202

130.1594

118.0862 622.0285 922.0101141.9593 307.0025 591.0169194.1175

+TOF MS: 3.773 to 4.124 min from 10ng_sulfa.wiff Agilent Max. 5.9e4 counts.

286 288 290 292 294 296 298 300 302 304 306 308 310 312m/z amu

0.0

1.0e4

2.0e4

3.0e4

4.0e4

5.0e4

5.9e4

Inte

nsity

, cou

nts

285.0202

287.0174

286.0228285.2858 307.0025288.0202

EIC of Target Adduct Ions

Average Mass Spectrum

Average Mass Spectrum --Target Mass Region Expanded

1

Figure 1 Automated empirical formula confirmation report for sulfachloropyridazine.

Visit our website at www.agilent.com/chem/tof to download the detailed note.

Detailed note: Automated empirical formula confirmation using the Agilent LC/MSD TOF, Agilent TechnologiesApplication Note, publication number 5989-0625EN (2004).

Figure 2 Chromatograms showing separation sulfa drugs on each of the alternating columns. A) sulfamethizole, B) sulfamethazine, C) sulfachloropyridazine, D) sulfadimethoxine

28

system by a system administratorand if empirical formula confirma-tion is desired, the target formulais input. The submitter is prompt-ed where to place the samples inthe autosampler and the systemindicates when they will be com-pleted. The samples are loggedinto a queue and when analyzed,

Technique Applications

Using accurate mass LC/MS in a walk-up environment

Instruments providing accurate

mass have been used in the devel-

opment of pharmaceuticals both

for confirming the identity of syn-

thesized compounds and as a pow-

erful tool in the identification of

unknowns. However, obtaining the

needed mass accuracy has required

the samples to be run by highly

skilled mass spectrometrists. In

todays pharmaceutical develop-

ment laboratories there is a desire

for samples to be run by the

chemists and not by dedicated

operators. The Agilent LC/MSD

TOF is capable of providing accu-

rate mass routinely. By having a

calibrant delivery system to auto-

matically dispense calibrant and a

solution for doing automated inter-

nal reference mass correction, any

operator skilled enough to run a

single quadrupole instrument can

obtain accurate mass spectra.

To make the operation even sim-pler, the Agilent Easy-AccessSoftware used on the LC/MSDQuadrupole was adapted for useon the LC/MSD TOF. To the sam-ple submitter, the user interface isvirtually indistinguishable fromthe quadrupole version. The sub-mitter merely logs in some sampleinformation, selects from somemethods previously set up on the

a report is printed out or e-mailedto the submitter. The user inter-face is shown in figure 1. If aspecified time has passed sincethe last sample was run, the sys-tem will automatically introducethe calibrant mixture and recalcu-late the time-to-mass coefficientsbefore running the next samples.

Figure 1The Easy-Access user interface showing status and for submitting samples.

29

Detailed note: Using accurate mass LC/MS in a walk-up environment, Agilent Technologies Application Note, publication number 5989-2548EN (2005).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

A variety of automated reportscan be produced by the system foreither confirming the synthesis oflead compounds or aid in identify-ing unknowns. The user canchoose from methods that printout the major ions for each peakin a chromatographic signal orpropose empirical formulas foreach of those ions. If a compounddatabase is available, the systemwill identify if any of the compoundsin the database are present. Forthe confirmation of correct syn-thesis of recombinant proteins,the submitter inputs a molecularweight or sequence and the systemoutputs a protein confirmationreport. An example of e-mailedreporting is shown in figure 2.

The system can be set up by a sys-tem administrator to match mostlaboratory protocols in areas suchas security, terminology, method-ology, etc. Changing the setup isunder password control. Theadministrator can even choose toreceive e-mail notification of sys-tem errors or low solvent state.The system can be set up to workwith submitters with a range ofskill levels. For example, someusers may be able to specify

Figure 2Example of an email notification of sample completion with atttached report.

injection volume, move samples inthe queue and have a wider rangeof methods while others may havemuch more limited capability. Inlaboratories where tracking sam-ples for accounting purposes isneeded, the system will report theusage to a file or database.

The Easy-Access software allowslaboratories to obtain resultsrequiring accurate mass measure-ment without requiring highlyskilled instrument operators.

Automated empirical formula generation for the identification of unknown compounds

30

Mass spectrometers that provide

accurate mass measurement are a

powerful tool in the identification

of unknown compounds. Modern

LC/MS instruments with Time of

Flight analyzers can often provide

mass measurements with errors

under 3 ppm. While this cannot

usually provide a unique empirical

formula for a complete unknown,

in many cases such as the analysis

of impurities in the chemical syn-

thesis of lead molecules in pharma-

ceutical development, this mass

accuracy can be used with other

information to provide a single log-

ical empirical formula. This note

describes how the Agilent LC/MSD

TOF can be used to provide auto-

mated reporting of empirical for-

mulas for the peaks in a HPLC

chromatographic signal.

Experimental

The Agilent LC/MSD TOF is capa-ble of providing routine accuratemass measurements with minimaluser interaction. A calibrant deliv-ery system (CDS) automaticallyprovides both a calibrant mix forthe tuning and calibration of theinstrument but also a referencemix through a second nebulizer fordoing real time internal referencecalibration1. Earlier versions of the

software provided an automatedMass List Report that provided themass spectra from each peak in aTotal Ion Chromatogram (TIC).This method was modified toexpand the allowed signals toinclude a Base Peak Chromato-gram (BPC), UV signal or the ana-log signal of any other detectorsuch as a light scattering detector.Additionally, the user can selectempirical formula generation andspecify a list of elements to con-sider with minimum and maximumvalues for each. The electron stateand charge state are also specified.The user can select a mass errorrange to consider and a maximumnumber of hits. The results report-ed can be sorted by mass error orby considering the isotope values.In this case, the isotopes of theadduct ion cluster are also consid-ered. For example, they can becompared to the theoretical iso-tope ratio and a score out of 1000reported2. In most cases, the cor-rect formula will have a score of990 or greater. Figure 1 containsan example of the first page of areport based on the BPC of a mix-ture of 4 sulfa drugs. In this exam-ple, the top formula on the list,sorted by mass error, is the correctone.

Conclusion

Using the automated generation ofempirical formula, the AgilentLC/MSD TOF is a powerful tool inthe identification of unknownswhile still being easy to use.

Technique Applications

31

Detailed note: Automated empirical formula generation for the identification of unknown compounds, AgilentTechnologies Application Note, publication number 5989-2779EN (Summer 2005).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

Figure 1First page of an Empirical Formula Generation report.

32

Technique Applications

Two approaches are used for doingthis form of target screening orcompound confirmation. The firstis to extend the standard mass listreport that takes each peak in thechromatographic signal andobtains a mass spectrum. This sig-nal can be the Total ion chro-matogram (TIC), base peak chro-matogram (BPC), UV trace or theanalog signal from another detec-tor. The user selects possibleadducts, charge states, etc to con-sider and how many masses ineach spectrum to consider. Foreach ion, based on the adduct list,possible molecular weights are cal-culated. These molecular weightsare compared against the entriesin the database. The database is acomma separated value (CSV) filecontaining the following informa-tion for each compound:

Formula, Retention Time,Molecular Weight, CompoundName, Description

The retention time is an optionalvalue that allows for excludingentries that do not fall within aspecified window.

The second approach is to do areverse search and is an extensionof the empirical formula confirma-tion (EFC) report described in aprevious note1. Instead of specify-ing a specific empirical formula toconfirm, the user specifies a data-base and for each formula in thedatabase, considers the selected

adduct ions and generates anextracted ion chromatogram(EIC). An average spectrum isobtained if the EIC is of sufficientabundance and the ions in thisspectrum are compared to the cal-culated m/z values. This type ofsearch is very useful when lookingfor targets in complex matricesand the compounds do not showup as peaks in any of the possiblesignals.

As an example of this type ofscreen, a mixture of 14 streetdrugs2 were searched against adatabase of 50 pharmaceutical anddrug of abuse compounds. Thesearch was based on the BPC fromm/z 125 to 600. The first page ofthe report can be seen in figure 1.By considering retention time, theisomers hydrocodone and codeineare correctly identified.

By using a database of knownempirical formulas, the accuratemass capabilities of the LC/MSDTOF allow for confident identifica-tion of target compounds inunknown samples.

Screening for target compounds usingthe LC/MSD TOF

Mass spectrometry has long been

used to identify compounds using

database searching. In GC/MS,

databases of hundreds of thou-

sands of compounds exist. These

databases contain the electron ion-

ization spectra of (EI) each com-

pound in the database, typically

reduced to 10 to 25 nominal mass

values. The unknown spectrum is

compared to the entries in the

database and the hits are ranked

using one of a variety of algorithms.

In LC/MS a compound typically

produces far fewer ions and the

ones obtained vary greatly depend-

ing on the HPLC conditions used.

The MS/MS spectrum can be used

as it typically contains ions other

than those indicating molecular

weight. However, rather than

identifying a compound based on

the presence of many ions at nomi-

nal mass, another approach is to

use simply an isotope cluster

indicative of the molecular weight

but using accurate mass measure-

ment. In this case, the database

need contain no spectral information

at all, merely an empirical formula

and the calculated monoisotopic

weight. This note describes how

this can be done with the Agilent

LC/MSD TOF, an instrument capa-

ble of routinely providing mass

measurements with better than

3 ppm mass accuracy.

33

Detailed note: Screening for target compounds using the LC/MSD TOF, Agilent Technologies Application Note, publication number 5989-2780EN, (Summer 2005).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

Figure 1First page of an Empirical Formula Generation report.

34

Technique Applications

Fast scanning for high throughputscreening using the LC/MSD TOF

In pharmaceutical drug discovery,

the large number of lead molecules

being analyzed has put pressure on

investigators to run increasing

number of samples in less time

than in the past. The use of short,

small particle size columns run at

relatively high flow rates now

results in chromatographic run

times of less than a minute. The

resulting peaks elute with baseline

peak widths of less than 2 seconds.

This high-speed chromatography

places demands on the total HPLC

system. In particular, it is impor-

tant that the detector acquisition

rate produces enough data points

across the peak without sacrificing

the quality of the information. For

mass spectrometer detectors, the

traditional 1-2 spectra per second

rate is no longer suitable.

The time an ion takes to passthrough a quadrupole mass analyz-er limits the possible scan speedunless resolution is sacrificed. Fora time-of-flight analyzer, a singletransient or scan takes place in 70 to 100 microseconds. At normalscan speeds, 5,000 10,000 tran-

sients are summed to get a singlespectrum. If faster scan speeds aredesired, one merely sums up fewertransients. Of course, at fasteracquisition rates, the data process-ing requirements are much moredemanding. The LC/MSD TOF iscapable of acquiring spectra over arange of m/z 100 1000 at the rateof 20 per second or better, with noloss in resolution or mass accura-cy. Because fewer transients aresummed, the data will be noisier.

The LC/MSD TOF had the HPLCmodified for high throughput asdescribed in a previous note. 4.6 x 50 mm x 1.8 micron SB C-18columns were used at 2 ml/min.Because of the high flow, a simplesplitter using 2 5 cm pieces of0.005 ID PEEK tubing and a teereduced the flow to MSD to about1 mL/min. This yielded peaksapproximately 2 seconds wide atbaseline. In order to obtain astrong ion indicative of molecularweight as well as fragment ions forstructural information, a dualexperiment analysis was per-formed. The fragmentor was tog-gled from 175 volts to 250 volts

10 times per second. This gave atotal of 20 spectra per second or10 per second at each voltage.Mass accuracy is still maintainedand sufficient data points areobtained across even a 2 secondwide peak.

The chromatogram obtained for afour-component mixture of sulfadrugs is shown in figure 1.Representative spectra are shownfor both the low and high fragmen-tor voltages.

35

Detailed Note: Fast scanning for high throughput screening using the LC/MSD TOF, Agilent TechnologiesApplication Note, publication number 5989-2558EN, (2005).

Visit our website at www.agilent.com/chem/tof to download the detailed note.

Base peak chromatogram of +TOF MS: dual frag at 10 Hz

0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90Time (min)

0.0

4000.0

8000.0

1.2e4

1.6e4

2.0e4

2.4e4

2.8e40.81

0.63

0.57

0.68

0.36

Intensity, cps

A

B C

Fragmentor 175 volts

Max. 1.4e4 counts.

120 160 200 240 280 320 360 400 440 480m/z, amu

120 160 200 240 280 320 360 400 440 480m/z, amu

0%10%20%30%40%50%60%70%80%90%

100%

Rel. Int. (%)

0%10%20%30%40%50%60%70%80%90%

100%

Rel. Int. (%)

311.0814

Mass error 1.757 ppm

Fragmentor 250 volts

Max. 1.1e4 counts.311.0805

333.0630156.0754

Mass error 1.136 ppm

Figure 1A) Base peak chromatogram of 4 sulfa drugs.B) Spectrum of sulfadimethoxine at 175 volts. C) Spectrum of sulfadimethoxine at 250 volts.

Structure elucidation of degradationproducts from the antibiotic drugamoxicillin by accurate mass measurement using LC-ESI-oaTOF MS

36

A final dosage from the antibiotic

drug amoxicillin was degraded

under induced stress conditions.

Herein we describe the identifica-

tion and confirmation of degrada-

tion products by structure elucida-

tion with accurate mass determina-

tion of the molecular ions and of

the CID fragments by LC-ESI-oa

TOF MS.

In modern pharmaceutical drugdiscovery and development it is ofcrucial importance to identify anunknown compound with thehighest possible degree of confi-dence because of its potentialtoxic effects on humans. The com-pound could be, for instance thepharmaceutical active substanceitself, a minor byproduct of theproduction process, a secondarysubstance in a drug isolated froma natural source, a metabolite cre-ated in the human body or adegradation product of the phar-maceutical agent created underharsh storage conditions. In addi-tion to the repertoire of analyticalmethods for structure elucidationa common method for the identifi-cation and identity confirmationof an unknown compound is themass spectrometric determinationof accurate molecular mass andconsequently the calculation ofthe empirical formula. Severalyears ago only operation intensivemagnetic sector field and FT massspectrometers were able to per-form these measurements withsufficient accuracy. Nowadays,with the advent of new TOF tech-nologies comparably easy to use

and inexpensive ESI orthogonalacceleration TOF instruments arealso capable of handling this task.Recently published examples ofan LC-ESI-TOF instrument used instructure elucidation are the iden-tification of a photo oxygenationproduct of a broad-spectrumantibiotic for livestock as well asthe identification of highly com-plex polyene macrolides isolatedfrom Streptomyces noursei.In this work the identification bystructure elucidation of degrada-tion products from the antibioticdrug amoxicillin obtained understress conditions with accuratemass determination of the molecu-lar ions and of the CID fragmentswith LC/ESI-oaTOF for the confir-mation of the molecular formulaand the fragment formulas isdescribed.

Experimental

ESI-oaTOF instrument: AgilentLC/MSD TOF equipped with a dualsprayer source (positive mode) forthe simultaneous infusion of thelock mass reference solution. Dry gas: 7.0 L/min Dry temp.: 300 C Nebulizer: 15 psi Scan: 50-1000 m/zFragmentor: 150 V or 300 V for CIDSkimmer: 60 VCapillary: 5000 V

LC system: Agilent 1100 Series capillary LC system containing a capillary pump with micro vacuumdegasser, a micro well-plateautosampler with thermostat and a column compartment.Solvent A: Water, 10 mM

ammonium formate, pH 4.1

Solvent B: AcNColumn flow: 8 L/min Primary flow: 500-800 L/minGradient: 0 min 0 % B, 1 min

0 % B, 13 min 25 % B,23 min 25 % B

Stop time: 23 min Post time: 15 minColumn: ZORBAX SB Aq

0.3 mm x 150 mm, 3.5 m

Sample preparation: To treat theantibiotic amoxicillin with acidicstress conditions a solution of amoxicillin (25 mg/L) in 0.1 M HCl was stirred for one hourat room temperature (RT = 25 C).

Results and discussion

The degradation of Amoxicillin (1)was induced by subjecting thepure drug substance to harsh con-ditions as described in the experi-mental section. Dilutions of thissolution were collected at varioustime frames and subjected to capil-lary chromatography to separatethe accumulated degradation prod-

Technique Applications

37

ucts. The obtained total ion chro-matogram (TIC) clearly shows thedegradation of amoxicillin intovarious products (figure 1). Theoutlined chromatogram alsoassigns also the degradation prod-ucts, which were identified bystructure elucidation with accuratemass determination using ESI-TOFMS for the confirmation of themolecular ions and the CID frag-ments. The measurement was per-formed twice using different frag-mentor voltages of 150 V and 300 Vin the TOF MS. When the fragmen-tor voltage is set at 150 V there isno collision induced dissociation(CID) observed whereas a goodfragmentation for the molecularion from the separated degrada-tion products is observed at a volt-age of 300 V. The first degradationproduct of amoxicillin (1) obtainedafter braking the four memberedbeta lactame ring amoxicillin peni-cilloic acid (2) was confirmed byaccurate mass determination withthe ESI-TOF MS with m/z 384.1212,with 4.50 ppm mass accuracy andformula calculation (insert in fig-ure 1). The structure of this degradation product was elucidat-ed by the appearance of specificfragments in the ESI-TOF mea-surement applying a fragmentorvoltage of 300 V with m/z 323.1063with 0.78 ppm mass accuracy,which is the product of a deamina-tion followed by a decarboxylationreaction (figure 2). The completefragmentation pattern of amoxi-cillin penicilloic acid (2) alsoshows the fragments with m/z 189.0697 and 122.0605 with arespective high mass accuracy of0.39 ppm and 0.70 ppm, which areimportant for the structure confir-mation (insert in figure 2).

2 4 6 8 10 12 14 16 18 20 22Time (min)

0.01.0e62.0e63.0e64.0e65.0e66.0e67.0e68.0e69.0e61.0e71.1e71.2e71.3e71.4e71.5e71.6e71.7e7

Intensity, cps

1

2

3

4

5

1. Amoxicillin2. Amoxicillin penicilloic acid3. Amoxicillin penilloic acid I and II4. Diketopiperazine amoxicillin 5. 4-Hydroxyphenylglyl amoxicillin

374.0 378.0 382.0 386.0 390.0 394.0m/z (amu)

0.02.0e46.0e41.0e51.4e51.8e52.2e52.6e53.0e53.4e53.8e54.2e5

Intensity, counts

384.1212

385.1235386.1239

Formula: C16H22N3O6SCalculated Mass: 384.1229Measured mass: 384.1212Mass accuracy: 1.70 mDa Mass accuracy: 4.50 ppm Fragmentor: 150 V

Figure 1Total ion chromatogram of TOF from amoxicillin (1) and its degradation products.Insert: Exact mass determination of amoxicillin penicilloic acid (2) (C16H21N3O6S),[M+H]=384.122 m/z.

2R=COOH

367.0963

323.1063122.0606 189.0697

160.0433

120 140 160 180 200 220 240 260 280 300 320 340 360 3800.00

1000.00

2000.00

3000.00

4000.00

5000.00

6000.00

7000.00

8000.00

9000.00

1.00e4

Intensity, counts

189.0697

384.1225

114.0371323.1063

160.0433367.0963

m/z (amu)

122.0605