separation science redefined - waters · ultra performance liquid chromatography coupled to...

ADVA

NST

ARH

AN

PU

BLI

CAT

ION

SUPPLEMENT TO

Ultra Performance LC™

Separation Science Redefined

Ultra Performance LC™


© 2004 Waters Corporation. Waters, Micromass, NuGenesis, Connections, ACQUITY Ultra Performance LC, ACQUITY UPLC, Alliance, LCT Premier, Quattro Premier, Q-Tof, ZQ, Connections AssetCARE, Connections INSIGHT,

Connections AQT, Atlantis, Oasis, Symmetry, Empower, MassLynx and eLab Notebook are trademarks of Waters Corporation.

From discovery to manufacturing, Waters has the tools you need to create a successful pharmaceutical product. Our ever-expanding range of technologies

and services – including sample preparation, separation, detection, information management and more – helps make your process more efficient, so your

products get to market faster. Add to that the dependability and accuracy you’ve come to expect from Waters and you’ll see that Waters is the only place

to turn for complete solutions – and for complete confidence. Learn more at www.waters.com/pharma

Everything you need tocreate a successful product.

Take as needed.

MAY 2005 SEPARATION SCIENCE REDEFINED 3www.chromatographyonline.com

4

6

8

15

22

31

36

40

45

Ultra Performance LCSeparation Science Redefined

Ultra Performance LCSeparation Science Redefined

Introduction — Separation Science RedefinedArthur G. Caputo

PrefaceMichael E. Swartz

Ultra Performance Liquid Chromatography(UPLC™): An IntroductionMichael E. Swartz

The Use of ACQUITY UPLC™ in PharmaceuticalDevelopmentAnton D. Jerkovich, Rosario LoBrutto, and Richard V. Vivilecchia

Ultra Performance Liquid Chromatography Coupledto Orthogonal Quadrupole TOF–MS(MS) forMetabolite IdentificationIain Beattie, Karine Joncour, and Kim Lawson

Assay Transfer from HPLC to UPLC for HigherAnalysis ThroughputYing Yang and Craig C. Hodges

The Evaluation and Application of UltraPerformance Liquid Chromatography (UPLC) for theRapid Analysis of Dose FormulationsSalane King, Peter J. Stoffolano, Eric Robinson, Thomas Eichhold, StevenH. Hoke, Timothy R. Baker, Eloise C. Richardson, and Kenneth R.Wehmeyer

Developing Columns for UPLC: DesignConsiderations and Recent DevelopmentsEric S. Grumbach, Thomas E. Wheat, Marianna Kele, and Jeffrey R. Mazzeo

Bibliography

4 SEPARATION SCIENCE REDEFINED MAY 2005 www.chromatographyonline.com

INTRODUCTIONINTRODUCTION

n 2004, Waters introduced a new category of LC technology that has changed separationscience forever.

We call it ACQUITY Ultra Performance LC or UPLC.

Based on novel chemistry and instrumentation, UPLC delivers new levels of resolution,speed and sensitivity.

I am pleased to report that interest in UPLC from scientists worldwide has been outstand-ing, from those looking to drive new products through the development process faster, to thoselooking for a more robust technique for routine analytics, to researchers who need to push MSperformance to the next level.

Scientists are now realizing they can do and see with UPLC what they couldn’t do and seewith HPLC. Some of their exciting work is found in the pages of this supplement. As you willsee, the benefits are real and they are compelling.

Throughout 2005, Waters will be adding to its existing portfolio of UPLC products. Asrecently as March, Waters introduced three new chemistries: a C8, phenyl, and a Shield RP18,as well as an evaporative light scattering detector, further expanding applications for UPLC.

Waters is committed to working with you to gain insight into your daily challenges and usethis understanding to direct our technology development efforts. We remain the only companythat can offer you a comprehensive set of technologies for UPLC with one goal in mind: to getyou quality information — faster.

UPLC is here to stay and we are confident it will continue to have a lasting impact in labsaround the globe.


Art CaputoPresident, Waters Division

Waters Corporation

I

© 2004 Waters Corporation. Waters is a trademark of Waters Corporation.

Today, we have more technologies and capabilities than ever before. From Chemistry and Chromatography to Informatics, Mass Spectrometry and Support Services,we have systems and solutions that are right for you. Solutions that will give you more answers. More time to look for them. More speed and accuracy. More efficiency and productivity. More service. More uptime and less downtime. More reliability and less impossibilities. We’re a partner who will work with you to understand your needs and fuel your success. A partner who can give you more confidence in your work than ever before. To learn more, visit www.waters.com

WE’RE A LOT MORE THAN YOU THINKSO YOU CAN DO A LOT MORE THAN YOU THOUGHT

WE’RE A LOT MORE THAN YOU THINKSO YOU CAN DO A LOT MORE THAN YOU THOUGHT

CHEMISTRY CHROMATOGRAPHY INFORMAT ICS MASS SPECTROMETRY SERV ICES


PREFPREFACEACE

e hope you find this special Ultra Performance Liquid Chromatography supplement toLCGC useful and informative. Each of the manuscripts presents a unique view and use

of this exciting new technology.The first article serves as an introduction, highlighting the theory and implementation of

UPLC, including the technological strides necessary in chemistry and instrumentation in orderto capitalize on UPLC’s increased speed, sensitivity, and resolution. The article goes on todescribe how UPLC can be used in both drug discovery (in-vitro metabolism) and environ-mental applications.

The next article describes the use of UPLC in pharmaceutical development. The benefits offast method development increasing sample throughput and laboratory productivity, and anexample of an eight-fold reduction in analysis time, without compromising resolution, arereported.

The third article further investigates drug discovery applications, with UPLC coupled toorthogonal quadrupole time-of-flight–mass spectrometry (TOF–MS). The authors show sig-nificant gains afforded by UPLC over conventional capillary-scale liquid chromatography–massspectrometry (LC–MS) for metabolite identification and MS spectral quality.

In the next article, an HPLC assay is converted and optimized for UPLC, achieving bothhigher sample analysis throughput and better assay sensitivity. A general strategy for methodconversion is summarized, and an analysis of operation costs and sample throughput foundUPLC superior to HPLC in the quality control (QC) laboratory.

The fifth article examines the use of 1 minute high-speed UPLC separations for dose formu-lation strength analysis. The chromatographic parameters evaluated include retention time andpeak reproducibility, as well as resolution and column ruggedness in isocratic and gradientseparations.

The final article provides details on the considerations taken when developing a new chro-matographic particle for UPLC separations. Information on column stability under aggressivetesting conditions is included. Additionally, information on several new column chemistries toprovide the utmost flexibility for methods development is reported.

Also included in this supplement is an up-to-date bibliography of published UPLC refer-ences the reader can consult for more information.

At a time when many scientists have reached separation barriers with conventional HPLC,UPLC presents the possibility to extend and expand the utility of chromatography and obtainquality results faster, redefining separation science.

Your comments and thoughts are always welcome.

W

Michael Swartz, Ph. D.Principal Scientist, Waters Corporation,

and LCGC Editorial Advisory [email protected]

When quantitative analysis presents difficult challenges, Waters is the obvious solution. Combining the revolutionary ACQUITY UPLCSystem with Waters’ Quattro family of mass spectrometers, you can achieve fast, reliable, highly sensitive LC/MS/MS analyses with easeRobust and reproducible results are assured with the Quattro micro™ and its ESCi™ multi-mode ionization capabilities. The Quattro Premieradds advanced T-Wave™ technology for unmatched speed, sensitivity and specificity. And data processing and method development tasks are simplified with dedicated software application managers, including QuanLynx™, QuanOptimize™ and TargetLynx™. With Waters all the tools for success are at your fingertips. To find out more about our quantitative analysis solutions, visit www.waters.com/quan

™

.™

,

All the tools to get every job done right.All the tools to get every job done right.

©2005 Waters Corporation Waters, ACQUITY UPLC, Quattro micro, ESCi, Quattro Premier, T-Wave, QuanLynx, QuanOptimize and TargetLynx are trademarks of Waters Corporation.


Ultra Performance Liquid Chromatography(UPLC): An Introduction

Michael E. Swartz, Ph.D. Principal Scientist, WatersCorporation, Milford,Massachusetts, [email protected].

igh performance liquid chro-matography (HPLC) is aproven technique that has

been used in laboratories worldwide over thepast 30-plus years. One of the primary driv-ers for the growth of this technique has beenthe evolution of packing materials used toeffect the separation. The underlying princi-ples of this evolution are governed by thevan Deemter equation, which is an empiri-cal formula that describes the relationshipbetween linear velocity (flow rate) and plateheight (HETP or column efficiency). Sinceparticle size is one of the variables, a vanDeemter curve can be used to investigatechromatographic performance.

According to the van Deemter equation,as the particle size decreases to less than 2.5 mm, not only is there a significant gainin efficiency, but the efficiency does notdiminish at increased flow rates or linearvelocities. By using smaller particles, speedand peak capacity (number of peaks resolvedper unit time in gradient separations) can beextended to new limits, termed Ultra Perfor-mance Liquid Chromatography, or UPLC.The technology takes full advantage of chro-matographic principles to run separationsusing columns packed with smaller particlesand/or higher flow rates for increased speed,with superior resolution and sensitivity.

Figure 1 shows a stability indicating assayof five related substances accomplished inunder one minute, proving that the resolv-ing power of UPLC is not compromisedeven at high speed. The current USP lists

multiple HPLC methods for the analysis ofthese same compounds with run timesapproaching 20 min, with broad, tailedpeaks.

Chemistry of Small ParticlesAs shown in Figure 1, smaller particles pro-vide not only increased efficiency, but alsothe ability to work at increased linear veloc-ity without a loss of efficiency, providingboth resolution and speed. Efficiency is theprimary separation parameter behind UPLCsince it relies on the same selectivity andretentivity as HPLC. In the fundamentalresolution (Rs) equation:

resolution is proportional to the squareroot of N. But since N is inversely propor-tional to particle size (dp):

as the particle size is lowered by a factor ofthree, from, for example, 5 mm (HPLC-scale) to 1.7 mm (UPLC-scale), N isincreased by three and resolution by thesquare root of three or 1.7. N is alsoinversely proportional to the square of thepeak width:

ACQUITY Ultra Performance LC™ Systems take advantage of technologicalstrides made in particle chemistry performance, system optimization,detector design, and data processing and control. When taken together,these achievements have created a step-function improvement inchromatographic performance. Defined as UPLC™, this new category ofanalytical separation science retains the practicality and principles of HPLCwhile increasing the overall interlaced attributes of speed, sensitivity, andresolution.

H

=N a21 k 4 a k 1 1( ) )(Rs 5System Selectivity

EfficiencyRetentivity

dp1

N }


This illustrates that the narrower thepeaks are, the easier they are to separate fromeach other. Also, peak height is inverselyproportional to the peak width:

So as the particle size decreases to increaseN and subsequently Rs, an increase in sensi-tivity is obtained, since narrower peaks aretaller peaks. Narrower peaks also mean morepeak capacity per unit time in gradient sep-

arations, desirable for many applications,e.g., peptide maps.

Still another equation comes into playwhen migrating toward smaller particles:

0.1A

bso

rban

ce (

AU

)

Time (min)1.00.0

1

5

4

3

2

Figure 1: UPLC stability indicating assay. UPLC conditions: Column: 2.1 3 30 mm 1.7µm ACQUITY BEH C18 at 30 °C. A 45 s, 5–85%B linear gradi-ent, at a flow rate of 0.8 mL/min was used. Mobile phase A was 10 mm ammonium formate, pH 4.0, B was acetonitrile. UV detection at 273 nm and40 pts/s. Peaks are in order: 5-nitroso-2,4,6-triaminopyrimidine, 4-amino-6-chloro-1,3-benzenesulfanamide, hydrochlorthiazide, triamterine, andmethylbenzenesulfanamide; 5 mL injection, 0.1 mg/mL each.

1

Ab

sorb

ance

(A

U)

0.6

1.2

Time (min)

3

2

0.0

Figure 2: Conversion of an HPLC method to an UPLC method. Column: 2.1 3 50 mm, 1.7 mm ACQUITY BEH C18 at 40 °C. A 10 mM ammonium phos-phate (pH 9.2) acetonitrile linear gradient from 10–90% acetonitrile at a flow rate of 1.0 mL/min was used, with UV detection at 210 nm. Samplewas a 3 mL injection of a 5 mg/mL mixture of each of the compounds. Peaks are in order: pseudoephedrine, ibuprofen, and butylparaben internalstandard.

w2

1N } w

1H }

dpdpppFopt }

1


This relationship also is revealed from thevan Deemter plot. As particle size decreases,the optimum flow Fopt to reach maximumN increases. But since back pressure is pro-portional to flow rate, smaller particle sizesrequire much higher operating pressures,and a system properly designed to capitalizeon the efficiency gains; A system that canboth reliably deliver the requisite pressuresand that can maintain the separation effi-ciency of the small particles with tightlymanaged volumes.

Higher resolution and efficiency can beleveraged even further, however, when analy-sis speed is the primary objective. Efficiencyis proportional to column length andinversely proportional to the particle size:

Therefore, the column can be shortenedby the same factor as the particle size with-out loss of resolution. Using a flow rate threetimes higher due to the smaller particles andshortening the column by one third (againdue to the smaller particles), the separationis completed in 1/9 the time while maintain-ing resolution. So if speed, throughput, orsample capacity is a concern, theory can befurther leveraged to get much higherthroughput.

But the design and development of sub-2 mm particles is a significant challenge, andresearchers have been active in this area forsome time, trying to capitalize on theiradvantages (2–4). Although high efficiency,

nonporous 1.5-mm particles are commer-cially available, they suffer from poor load-ing capacity and retention due to low surfacearea. Silica-based particles have goodmechanical strength but can suffer from anumber of disadvantages, which include alimited pH range and tailing of basic ana-lytes. Polymeric columns can overcome pHlimitations, but they have their own issuesincluding low efficiencies, limited loadingcapacities, and poor mechanical strength.

In 2000, Waters introduced XTerra®, afirst generation hybrid chemistry that tookadvantage of the best of both the silica andpolymeric column worlds. XTerra columnsare mechanically strong, with high effi-ciency, and operate over an extended pHrange. They are produced using a classicalsol-gel synthesis that incorporates carbon inthe form of methyl groups. But in order to

dpL

N }

Figure 3: Full scan Tof MS for HPLC (Figure 3a) and UPLC analyses (Figure 3b) of the metabolites of dextromethorphan. (Reproduced with permis-sion from reference 12, copyright John Wiley and Sons Limited 2005.)

100

Time (min)

%

0

100

%

01.00 2.00 3.00 4.00 5.00

1.00 2.00 3.00 4.00 5.00

0.39

1.60

2.705.44

5.17

4.87

5.66

4.38

4.69

3.63

4.332.09

1.40

2.281.29

3.652.48 5.13

1.14

1.08

4.653.54

2.82

(a)

(b)


provide the necessary mechanical stabilityfor UPLC, a second generation bridgedethyl hybrid (BEH) technology was devel-oped. Called ACQUITY BEH, these1.7-mm particles derive their enhancedmechanical stability by bridging the methylgroups in the silica matrix.

Packing 1.7-mm particles into repro-ducible and rugged columns was also a chal-lenge that needed to be overcome. Require-ments include a smoother interior surface ofthe column hardware and re-designing theend frits to retain the small particles andresist clogging. Packed bed uniformity is alsocritical, especially if shorter columns are tomaintain resolution while accomplishing thegoal of faster separations. All ACQUITYUPLC BEH columns also include eCord™microchip technology that captures themanufacturing information for each col-

umn, including the quality control tests andcertificates of analysis. When used in theWaters® ACQUITY UPLC™ System, theeCord database is also updated with realtime method information, such as the num-ber of injections, or pressure and tempera-ture information, to maintain a completecontinuous column history.

Capitalizing on Smaller ParticlesInstrument technology also had to keep paceto truly take advantage of the increasedspeed, superior resolution and sensitivityafforded by smaller particles. StandardHPLC technology simply doesn’t have thecapability to take full advantage of sub-2 mmparticles. A completely new system designwith advanced technology in the solvent andsample manager, auto sampler, detector, datasystem, and service diagnostics is required.

The ACQUITY UPLC System has beenholistically designed for low system anddwell volume to minimize dispersion andtake full advantage of small particle technology.

As alluded to previously, achieving smallparticle, high peak capacity separationsrequires a greater pressure range than thatachievable by today’s HPLC instrumenta-tion. The calculated pressure drop at theoptimum flow rate for maximum efficiencyacross a 15-cm long column packed with1.7-mm particles is approximately 15,000psi. Therefore, a pump capable of deliveringsolvent smoothly and reproducibly at thesepressures and that can compensate for sol-vent compressibility, while operating in boththe gradient and isocratic separation modesis required.

With 1.7-mm particles, half-height peak

100

%

01.00 2.00 3.00 4.00 5.00 6.00 7.00

1.00 2.00 3.00 4.00 5.00 6.00 7.00

100

%

0

0.70

3.53

6.91

6.786.43

5.90

3.12

5.503.07

5.074.86

4.66

4.273.87

3.332.38

2.27

1.86

1.73

1.56

1.031.21

1.11

1.09

0.98

0.14

3.64

3.13

2.40 2.49

3.10

Time (min)

m/z = 244

m/z = 258

Figure 4: Extracted ion chromatograms for major N and O dealkylated and double de-alkylation metabolites of dextromethorphan by HPLC/TofMS. (Reproduced with permission from reference 12, copyright John Wiley and Sons Limited 2005.)


widths of less than one second can beobtained, posing significant challenges forthe detector. In order to accurately andreproducibly integrate an analyte peak, thedetector sampling rate must be high enoughto capture enough data points across thepeak. In addition, the detector cell musthave minimal dispersion (volume) to pre-serve separation efficiency. Conceptually, thesensitivity increase for UPLC detectionshould be 2–3 times higher than HPLC sep-arations, depending on the detection tech-nique. MS detection is significantlyenhanced by UPLC; increased peak concen-trations with reduced chromatographic dis-persion at lower flow rates (no flow splitting)promotes increased source ionization effi-ciencies (reduced ion suppression) forimproved sensitivity.

Sample introduction is also critical. Con-

ventional injection valves, either automatedor manual, are not designed and hardened towork at extreme pressure. To protect the col-umn from experiencing extreme pressurefluctuations, the injection process must berelatively pulse-free. The swept volume ofthe device also needs to be minimal toreduce potential band spreading. A fastinjection cycle time is needed to fully capi-talize on the speed afforded by UPLC,which in turn requires a high sample capac-ity. Low volume injections with minimalcarryover are also required to realize theincreased sensitivity benefits.

The ACQUITY UPLC System consists ofa binary solvent manager, sample manager(including the column heater), detector, andoptional sample organizer. The binary sol-vent manager uses two individual serial flowpumps to deliver a parallel binary gradient

mixed under high pressure. There are built-in solvent degassing as well as solvent selectvalves to choose from up to four solvents.There is a 15,000 psi pressure limit (about1000 bar) to take full advantage of the sub-2-mm particles. The sample manager alsoincorporates several technology advance-ments. Low dispersion is maintainedthrough the injection process using pressureassist sample introduction, and a series ofpressure transducers facilitate self monitor-ing and diagnostics. It uses needle-in-needlesampling for improved ruggedness and aneedle calibration sensor increases accuracy.Injection cycle time is 25 s without a washand 60 s with a dual wash used to furtherdecrease carry over. A variety of microtiterplate formats (deep-well, mid-height, orvials) can also be accommodated in a ther-mostatically controlled environment. Using

Figure 5: Extracted ion chromatograms for major N and O dealkylated and double de-alkylation metabolites of Dextromethorphan by UPLC/TofMS. (Reproduced with permission from reference 12, copyright John Wiley and Sons Limited 2005.)

100

%

01.00 2.00 3.00 4.00 5.00

1.00 2.00 3.00 4.00 5.00

100

%

0

Time (min)

0.40

m/z = 244

m/z = 258

5.70

4.49 5.404.26 5.114.86

4.04

2.08

2.09

2.67


the optional sample organizer, the samplemanager can inject from samples from up to22 microtiter plates. The sample manageralso controls the column heater. Columntemperatures up to 65 °C can be attained. A“pivot out” design provides versatility toallow the column outlet to be placed incloser proximity to the source inlet of an MSdetector to minimize excess tubing and sam-ple dispersion.

The tunable UV–vis and PDA detector sinclude new electronics and firmware tosupport Ethernet communications at thehigh data rates necessary for UPLC detec-tion. Conventional absorbance-based opticaldetectors are concentration-sensitive detec-tors, and for UPLC, the flow cell volumewould have to be reduced in standardUV–vis detectors to maintain concentrationand signal. However, smaller volume con-ventional flow cells would also reduce thepath length upon which the signal strengthdepends; and worse: a reduction in cross-sec-tion means the light path is reduced, andtransmission drops, increasing noise. There-fore, if a conventional HPLC flow cell isused, UPLC sensitivity would be compro-mised. The ACQUITY UPLC System

detector cell consists of a light guided flowcell equivalent to an optical fiber. Light isefficiently transferred down the flow cell inan internal reflectance mode that still main-tains a 10-mm flow cell path length with avolume of only 500 nL. Tubing and connec-

tions in the system are efficiently routed tomaintain low dispersion and to take advan-tage of leak detectors that interact with thesoftware to alert the user to potentialproblems.

ApplicationsScientists are used to making compromises;and one of the most common scenariosinvolves sacrificing resolution for speed.With UPLC increased resolution in shorterrun times can generate more informationfaster without sacrifices.

Higher sample throughput with moreinformation per sample may decrease thetime to market, an important driving forcein today’s pharmaceutical industry.

Figure 2 illustrates a method convertedfrom HPLC to UPLC that takes advantageof the speed of UPLC. The correspondingHPLC separation takes in excess of 12 min;UPLC accomplishes the same separation inunder 30 s.

UPLC can also be used to significantlyimprove the success of the drug discoveryprocess. Drug discovery is heavily dependantupon the early prediction of metabolic fateand interactions of drug candidate mole-cules. To prevent “poor” candidates fromprogressing through the discovery process,factors such as metabolic stability, toxicmetabolite production, p450 inhibition, andinduction are all routinely monitored. Bythe mid-1990s, high performance liquid

0.5 1.5 2.5 3.5 4.5 5.5 6.5

Time (min)

31% Methanol

60%Curve 8

112

1116

17

1415

1310

5 7

89

6

4

3

2

Figure 6: Explosives analysis using ACQUITY UPLC. Column: 2.1 3 100 mm 1.7 mm ACQUITY UPLC BEH C18. A water methanol gradient from 31–60%methanol was used as shown, at a flow rate of 0.5 mL/min, with UV detection at 254 nm. Sample was a 10 mL injection of a 10 ppm mixture of eachof the compounds. Peaks are in order: 2,6 diamino- 4 nitrotoluene, HMX, 2,4 diamino- 6 nitrotoluene, RDX, 1,3,5-trinitrobenzene, 1,2-dinitroben-zene, 1,3-dinitrobenzene, nitrobenzene, tetryl, 2,4,6-Trinitrotoluene, 2-Amino-4,6-dinitrotoluene, 4-Amino-2,6-dinitrotoluene, 2,4-dinitrotoluene,2,6-dinitrotoluene, 2-nitrotoluene, 4-nitrotoluene, 3-nitrotoluene.

Higher samplethroughput withmore information persample may decreasethe time to market,an important drivingforce in today’spharmaceuticalindustry.


chromatography (HPLC) directly coupledto mass spectrometry (MS) was in routineuse in drug metabolism laboratories forthese types of studies (5–12). Enhancedselectivity and sensitivity, and rapid, genericgradients made LC–MS the predominatetechnology for both quantitative and quali-tative analyses. However, with the ever-increasing numbers and diversity of com-pounds entering development, and thecomplex nature of the biological matricesbeing analyzed, new analytical proceduresand technology were required to keep pacewith the testing demands. Unexpected, reac-tive, or toxic metabolites must be identifiedas early as possible to reduce the very costlyattrition rate. This quest for more accuratedata meant improving the chromatographicresolution to obtain higher peak capacity,reducing the co-elution of metabolites, whileenhancing the sensitivity and decreasing ionsuppression in the MS.

The power of the ACQUITY UPLC Sys-tem when used in drug discovery can beillustrated by the analysis of the in-vitrometabolism of dextromethorphan. Dex-tromethorphan undergoes O-dealkylation intwo positions leading to three major phase Imetabolites. These products can be furthermetabolized via conjugation with glucuronicacid to form metabolites of masses MH1 5434 and 420. The data in Figure 3a and 3bshows the HPLC–MS and UPLC–MS sepa-rations, respectively, of the in-vitro incuba-tion of dextromethorphan with rat livermicrosomes. As shown, the chromato-graphic performance of the ACQUITYUPLC BEH 1.7-mm particles is significantlybetter than that produced by the 3.5 mmmaterial. The 1.7-mm material gives peaks ofwidth 4 s at the base, resulting in a peakcapacity of over 100, whereas with HPLCthe average peak width was 20 s at the basegiving a total peak capacity of just 20, result-ing in a 5-fold increase in the performanceof the UPLC system.

The extracted ion chromatogram m/z 5258 and m/z 5 244 for the HPLC/MSanalysis is shown in Figure 4.

In Figure 4, we can clearly see the two O-dealkylated metabolites of dextromethor-phan m/z 5 258, these two metabolites areresolved to about 80%, while the 244metabolite is barely visible. These results canbe compared to those obtained by UPLC,here we can see that the two 258 ions areclearly resolved and that the 244 ion is noweasily detected, as illustrated in Figure 5.

This data clearly illustrates the improved

resolution and sensitivity of the UPLC sys-tem. This extra resolution is particularlyimportant when analyzing isobaric com-pounds such as these dealkylated metabo-lites. By incorporating a more efficientUPLC separation into the MS there is lession suppression from competing com-pounds in the source and therefore more dis-creet ionization of the metabolites. Withoutthe resolution generated by UPLC it wouldbe possible to falsely assign the structure of ametabolite or miss a potential toxic moiety.The extra sensitivity produced by the UPLCsystem ensures more low concentrationmetabolites will be detected, helping to pre-vent potentially toxic compounds from pro-gressing further into the drug discoveryprocess. This added sensitivity is extremelyimportant when performing MS–MS exper-iments as it can make the difference betweenobtaining an interpretable spectrum or not.

Sensitivity, selectivity, and analysis time(sample throughput) are also some of thechallenges analysts face when analyzing envi-ronmental samples such as soil and water.

Explosives residues in soil or environmen-tal waters are of both forensic and environ-mental interest. These types of assays provechallenging because of the selectivity neededto resolve positional isomers. Typical HPLCanalyses require viscous, buffered mobilephases operated at high temperatures, andanalysis times exceeding 30 min.

Figure 6 shows the separation of a com-plex mixture of explosive compounds in lessthan seven minutes, with a much simpler,more robust mobile phase than that com-monly used in HPLC assays. The simplernonbuffered mobile phase also is ideal forMS detection if desired.

ConclusionACQUITY UPLC using 1.7-mm particlesand a properly holistically designed systemprovide significantly more resolution (infor-mation) while reducing run times, andimprove sensitivity for the analyses of manycompound types. At a time when many sci-entists have reached separation barriers withconventional HPLC, UPLC presents thepossibility to extend and expand the utilityof chromatography. New ACQUITY tech-nology in both chemistry and instrumenta-tion boosts productivity by providing moreinformation per unit of work as UPLC ful-fils the promise of increased resolution,speed, and sensitivity predicted for liquidchromatography.

AcknowledgmentsThe author would like to acknowledge thecontributions of the ACQUITY programteam at Waters, and also Andy Aubin,Robert Plumb, Jose Castro-Perez, JoeRomano, Mark Benvenuti, Ramesh Rao,Jim Krol, and James Willis for their contri-butions to this manuscript.

References(1) J.J. van Deemter, F.J Zuiderweg, and A.

Klinkenberg, Chem. Eng. Sci. 5 (1956), p. 271.

(2) A.D. Jerkovitch, J.S. Mellors, and J.W. Jorgen-

son, LCGC 21(7), 2003.

(3) N. Wu, J.A. Lippert, and M.L. Lee, J. Chromo-

togr., 911(1), 2001.

(4) M.E. Swartz, J. Liq. Chromatogr., in press.

(5) J.P. Allanson, R.A. Biddlecombe, A.E. Jones, S.

Pleasance, Rapid Commun. Mass Spectrom.

10(811), 1998.

(6) I. M. Mutton, Chromatographia 47, (1998), p.

291.

(7) J. Ayrton, G.J. Dear, W.J. Leavens, D.N. Mal-

lett, R.S. Plumb, M. Dickins., Rapid Commun

Mass Spectrom., 12(5), (1998), 217–224

(8) J. Ayrton, G.J. Dear, W.J. Leavens, D.N. Mallett

and R.S. Plumb, J. Chromatogr., B 709(2)

(1998), 243–254.

(9) M. Jermal and Y. Xia, Rapid Commun. Mass

Spectrom. 13 (1999), p. 97.

(10) R.S. Plumb, G.J. Dear, D.N. Mallett and J.Ayr-

ton, Rapid Commun. Mass Spectrom. 15 (2001),

986–993.

(11) M.K. Bayliss, D. Little, D.N. Mallett and R.S.

Plumb, Rapid Commun. Mass Spectrom. 14

(2000), 2039–2045.

(12) J. Castro-Perez, R. Plumb, J.H. Granger, I. Beat-

tie, K. Joncour and A. Wright, Rapid Commun.

Mass Spectrom. 19, 843–848 (2005). n

At a time when manyscientists havereached separationbarriers withconventional HPLC,UPLC presents thepossibility to extendand expand theutility ofchromatography.


Anton D. Jerkovich, RosarioLoBrutto, and Richard V.VivilecchiaPharmaceutical and AnalyticalDevelopment, NovartisPharmaceuticals Corporation,East Hanover, New Jersey, [email protected]

he approaches for fast LC methoddevelopment are varied. Methoddevelopment simulation software

[such as ACD™ (2), DryLab™ (3), orChromsword™ (4)] is a valuable tool foroptimizing and streamlining methods. Suchsoftware allows the chemist to increase theinformation obtained from a limited num-ber of runs and to predict the best possibleseparation conditions. Of course, the bestseparation achievable is restricted by theinherent performance limits of the instru-ments and columns in use.

Since the quality of the separation mustnot be sacrificed, methods that offer thegreatest resolution per unit of time aredesired. Therefore, high-resolution chro-matographic techniques must be consideredin the development of fast LC methods.Monolithic columns, which contain a poly-merized porous support structure, providelower flow resistances than conventionalparticle-packed columns (1,5,6). Thesecolumns can be operated at higher flowrates, although increases in efficiency aremoderate. Also, solvent consumption is con-siderably higher and can be an issue for cost-conscious laboratories. Elevated-tempera-ture chromatography also allows for highflow rates by lowering the viscosity of themobile phase, which significantly reducesthe column backpressure (7). Interestingpossibilities arise from the use of high tem-peratures, such as the use of temperaturegradients and purely aqueous mobile phases(8). However, due to the limited availability

of packing materials stable at temperatures. 100 °C, method development approachesusing elevated temperature chromatographyare not yet considered routine.

The most straightforward and acceptedway to improve performance and gain speedin HPLC columns has been to reduce theparticle diameter of the packing material (9).By proportionally reducing the stationaryphase particle diameter (dp) and the columnlength, separation efficiency is maintainedwhile analysis time is reduced. Thus, shortercolumns (150 mm or less) packed withmaterial in the 3-mm size range have nowbecome standard where longer columns (upto 250 mm) with 5-mm materials were onceused. However, because the pressurerequired to pump mobile phase through thecolumn is inversely proportional to thesquare of the particle diameter, the backpres-sures required for use of these small-particlecolumns becomes high. Column manufac-turers are now able to reliably produce parti-cles in the sub-2 mm range (10,11), and thispresents a challenge to the pressure limita-tions of conventional HPLC systems. Sincechromatographers generally should operateat or above the optimum flow velocity for agiven column, even extremely short columnswith these particles reach the system pressurelimits before their full benefits can be real-ized. Consequently, there has been muchinterest in the use of elevated pressures (.10,000 psi) in LC to take advantage of theseparation speed that these particles can provide.

T

The Use of ACQUITY UPLC inPharmaceutical Development

Now more than ever, the demands of the pharmaceutical industry requirecompanies to look for new ways to cut costs and shorten timelines in thedevelopment of drugs, while at the same time improving the quality of theirproducts. The analytical laboratory is no exception to this trend. The developmentof HPLC methods for assay/purity analysis of drugs and their related substances is atime-consuming process and is often a bottleneck in analytical labs (1). Separationsscientists are thus continually driven to develop LC methods with ever-shorteranalysis times. The benefits of faster analyses are clear: they allow for a greaternumber of analyses to be performed in a shorter amount of time, therebyincreasing sample throughput and lab productivity. In addition, as test experimentsare performed more quickly, the overall method development time is decreased.


Until recently, research in this area hasbeen restricted to a few academic labs,mostly using non-porous particles packed infused-silica capillaries (12–19). The use ofnonporous particles, however, has been lim-ited in the pharmaceutical industry due totheir low sample loading capacity. The needhas therefore existed for a commercially fea-sible system to be used with columns packedwith porous particles less than 2 mm indiameter.

The ACQUITY Ultra Performance LC™(UPLC™) from Waters Corporation (Mil-ford, Massachusetts) is the first commer-cially available system that addresses thechallenge of using elevated pressures andsub-2 mm particles, which makes it a partic-ularly attractive and promising tool for fastLC method development. The ACQUITYsystem is capable of pumping mobile phaseat pressures up to 15,000 psi and utilizescolumns that are packed with stationaryphase particles 1.7 mm in diameter. Thepacking material is a porous, silica-basedbridged ethyl hybrid (BEH) material. Abonded C18 stationary phase was used forthe work presented here, although other col-umn chemistries are available. More thanjust a new pump, the instrument approaches

the use of these small-particle columns fromthe standpoint of the total HPLC system—both the engineering challenges of operatingat high pressures and the high performanceexpected from such columns necessitates aredesigned injector, reduced system vol-umes, an increased detector sampling rate,and other improvements.

To be suitable for the analysis of pharma-ceutical development samples under goodmanufacturing practices (GMPs), the UPLCinstrument and columns must not onlydeliver on its promises for fast, high-resolu-tion separations, but do so reproducibly andwith the required sensitivity. To this end, anumber of fundamental chromatographiccharacteristics were evaluated. Some param-eters that are of concern to analyticalchemists involved with pharmaceutical

methods development are column perform-ance, system band-broadening, injectionand gradient precision, and sensitivity.Results of these investigations are presentedhere. Also, a Novartis test method was trans-ferred to UPLC to demonstrate the benefitsit brings to our existing and future applications.

Efficiency and Band-BroadeningThe performance of a column can be meas-ured in terms of the height equivalent to atheoretical plate (H), which can be calcu-lated from the column length L, and the col-umn efficiency, or theoretical plates N

N

LH =

[1]

13.00

12.00

11.00

10.00

9.00

8.00

7.00

6.00

5.00

4.00

3.000.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Linear velocity (mm/s)

Methyl-parabenEthyl-parabenPropyl-parabenButyl-paraben

Plat

e h

eig

ht

(mm

)

Figure 1: Van Deemter curves for a series of parabens on a 2.1 3 100 mm ACQUITY UPLC BEH C18 column. Conditions: 30/70 v/v acetonitrile/watermobile phase, ambient temperature.

Column dimensions (mm) Column volume (ml)

1.0 3 50 281.0 3 100 552.1 3 50 1202.1 3 100 240

Table I: Column volumes of various ACQUITY column dimensions, calculated using equation 4


The number of theoretical plates is calcu-lated from the equation

where tR is the analyte’s retention time,and s is the standard deviation of the peak.A lower plate height indicates a more effi-cient column. The van Deemter equationdescribes H in terms of its dependence onthe linear flow velocity (u):

where A, B, and C are the coefficients foreddy diffusion, longitudinal diffusion, andresistance to mass transfer, respectively. Theoptimum column performance occurs at theminimum of the curve generated by plottingH versus u. A minimum plate height ofabout twice the particle diameter is generallyexpected for an efficient column.

Figure 1 shows van Deemter curves gener-ated on a 2.1 3 100 mm ACQUITY UPLC

BEH C18 column with isocratic elution(40/60 v/v acetonitrile/water mobile phase)at flow rates ranging from 0.05–0.45mL/min (1400–12,000 psi). The samplewas a series of parabens plus uracil as a deadtime marker, each at a concentration of50 mg/mL, and the injection volume rangedfrom 0.5-2 mL, equal to a sample load of25–100 ng of each compound on the col-umn. Detection was performed at 254 nm.This column yielded a minimum plateheight of about 4 mm (250,000 plates/m)for butyl-paraben (k 5 8.7), and about5.1 mm (200,000 plates/m) for methyl-

2

=s

RtN [2]

Cuu

BAH ++= [3]

100

90

80

70

60

50

40

30

20

10

0

0.0 2.0 4.0 6.0 8.0 10.0

Retention factor (k)

1.0 x 50 mm

1.0 x 100 mm

2.1 x 50 mm

2.1 x 100 mm

% o

f to

tal p

eak

varl

ance

Figure 2: The extracolumn contribution to band-broadening for various column dimensions. Conditions: 30/70 v/v acetonitrile/water mobile phase,ambient temperature.

Injection volume (ml) 20-ml injection loop 10-ml injection loop 5-ml injection loop

Pressure-assisted Pressure-assisted Pressure-assisted Needle-overfill

Phenol Toluene Phenol Toluene Phenol Toluene Phenol Toluene

0.5 3.2 3.3 3.1 4.0 3.0 3.2 1.2 2.11.0 0.5 1.0 0.9 1.4 2.2 2.0 0.1 0.62.0 0.8 1.6 0.9 0.8 0.8 1.2 1.1 0.75.0 0.4 0.9 0.5 0.7 0.3* 0.1* — —

* Full-loop injectionsFor fastest cycle times, Waters recommends pressure-assisted injection.For best injector precision, Waters recommends full-loop injection or partial-loop needle overfill with an injection volume of 30–70% of the fullloop volume.

Table II: Injection precision for various injection volumes and partial loop injection techniques on loop sizes of 5, 10, and 20 mL. Values cal-culated as %RSD of peak area, n 5 6. Retention factors for phenol and toluene are 1.3 and 12.7, respectively.


paraben (k 5 1.3). In terms of relative per-formance, this is equivalent to about 2.3 and3 times the particle diameter, respectively—slightly worse than what is theoreticallyexpected for particles of this size, althoughstill on par with most typical commercialpacking materials. The absolute efficiencyof this column (about 25,000 theoreticalplates) is only marginally better than what a3.0 3 150 mm, 3-mm particle column typ-ically provides. The speed at which this effi-ciency can be attained, however, is muchgreater, showing that speed of analysis is theACQUITY’s greatest benefit. The higherplate height (lower efficiency) for methyl-paraben is most likely due to extracolumnsources of band-broadening, which are morepronounced for earlier-eluting components.

This raises an important concern: thesmall dimensions and high efficiencies of thecolumns make extra-column band broaden-ing more of a concern than with typical ana-lytical columns (3-5 mm particles, 3–4.6mm i.d.), where extracolumn broadening isproportionally small enough that it does notsignificantly add to the overall broadening ofthe peak. Although system volumes in theACQUITY have been drastically reduced(total system volume , 100 mL) comparedto conventional HPLC systems (up to 1 mL), extra-column broadening is still evi-dent. Indeed, van Deemter analysis of a1.0 3 50 mm column demonstrated highplate heights of . 6 mm for butyl-parabenand . 12 mm for the earlier-eluting methyl-paraben.

Our experience has shown us that a goodrule of thumb is the post-injection extracol-umn volume of the system should notexceed about 10% of the column void vol-ume VM calculated as

VM ≈ 0.7* pr2 L [4]

where r is the column radius, L is the col-umn length, and 0.7 is the approximate frac-tion of the column occupied by mobilephase assuming porous particles (20). Thecolumn volumes for various ACQUITY col-umn formats are listed in Table I. The post-injection extracolumn volume can beapproximated by replacing the column witha “zero dead volume” union, injecting a testanalyte at a low flow rate, and calculating the

volume from the flow rate and the elutiontime. This method, while not exact, can givea general order of magnitude of the extra-column volume. The particular instrumentused for this work had a 5-mL injection loopand the low-flow “50/50” detector flow cell,which uses 50-mm capillaries at the inlet and

outlet and has a total volume of 500 nL.Nine measurements at various flow ratesresulted in values between 8 and 11 mL.This volume, while very small compared toconventional HPLCs, is large enough tocause significant band spreading in the 1.0-mm-i.d. formats.

The extracolumn contribution to the totalband spreading can also be determined fromthese measurements. Any broadening accu-mulated by the test analyte in this setup willbe due solely to non-column sources such asthe injector, connection tubing, and detec-tor flow cell. This of course assumes that theanalyte zone begins as an infinitely narrowband, which in actuality is not the case.Therefore, the smallest injection volumepossible—0.1 mL—was used. Because vari-ances are additive it is the peak variance,rather than peak width, that is representativeof individual contributions to bandbroadening:

The variance can be calculated from thechromatogram by the equation:

where W4.4% is the peak width at 4.4%peak height. Variances were measured atflow rates of 10 to 150 mL/min, beyondwhich elution times were too short toachieve reliable measurements. Since weknow from the Golay equation (21) that thevariance of a zone in an open tube increaseslinearly with the flow velocity, a least-squaresregression line was fit to the data and extrap-olated to higher flow rates to estimate extra-column variance contributions at those con-ditions. By comparing variances obtained inthis manner with that of peaks retained on acolumn, the impact of extra-column broad-ening can be assessed.

The same sample and conditionsdescribed above for the van Deemter analy-ses were used to obtain chromatograms atflow rates of 0.1 and 0.4 mL/min for the1.0-mm and 2.1-mm diameter columns,respectively. This corresponds to a linearvelocity of approximately 3 mm/s for each

222col.extra-columntotal += sss

[5]

2

%4.42

5

=W

s [6]

The retention timerepeatability for 200injections was <0.25%RSD for all 12peaks, indicating thatno drift in thegradient occurredthroughout theanalysis

Component Retention Time %RSD Ret. time (n 5 200)

1 0.50 0.252 0.57 0.233 0.65 0.204 1.18 0.14 5 1.72 0.186 1.82 0.207 1.87 0.178 1.97 0.219 2.06 0.1910 2.25 0.1311 2.36 0.0912 2.59 0.06

Table III: Retention time precision for 12 components separated on the ACQUITY UPLC. Atotal of 200 injections was performed over a period of more than 12 h. The run time was3.5 min.


column, which is near the optimum velocity.Variances were measured for methyl-,ethyl-, propyl-, and butyl-paraben, whoseretention factors (k), measured using uracilas a dead-time marker, ranged from about 1to 9. The extracolumn variance at the corre-sponding flow rate is then compared to thepeak variance. The contribution of extracol-umn broadening to the total broadening ofthe peak can be expressed as a percentage,and is plotted versus the retention factor forfour different column dimensions in Figure2. Two trends discussed earlier are made evi-dent from this plot: 1) extracolumn broad-ening has a greater effect on poorly retainedcomponents, and in fact becomes the domi-nant contributor to peak variance at low kvalues and 2) extracolumn broadening ismore pronounced the smaller the columnvolume is, so that it approaches 100% ofpeak broadening for 1.0-mm-i.d. columns.

Often, chromatographers will develop iso-cratic methods for fast assay determination,such as for dissolution or content uniformitytesting, or will incorporate an isocraticregion at the beginning of a gradient methodto retain the more polar constituents of asample. While the ACQUITY columns offergreater efficiency and speed, the minimumpeak width attainable under isocratic condi-tions will be limited by extra-column broad-ening, particularly with the 1.0-mm-i.d.columns. Gradient elution presents less of aproblem. Typically, sample components willbe concentrated onto the head of the col-umn by weak mobile phase starting condi-tions, reducing extracolumn effects thatoccur before the sample reaches the column.Early eluters, however, may still be affected.Therefore, it is hard to justify the use of the1.0-mm columns in the pharmaceuticaldevelopment environment, where sampleavailability is rarely an issue. These columnswill find their greatest use in labs where sam-ple amount is limited and high sensitivity isrequired, and for applications using massspectrometric detection. The 2.1-mmcolumns, however, exhibit little extra-col-umn broadening for components with k .5, and are quite suitable for routine methodsdevelopment.

Injection and GradientRepeatabilityAnother performance characteristic essentialfor a validated HPLC test method is injec-tion precision. The ACQUITY employs anovel injector design, and it was not knownhow its unique features would affect the

injector’s precision. This is not the place todiscuss the details of the injector design, buta brief description is warranted. Injectionscan be performed in a full loop mode or inone of two partial loop modes: pressure-assisted or needle-overfill. For full loopinjections, the loop is simply overfilled withsample. The injection valve is then switched,placing the loop in-line with the column.The pressure-assisted partial loop mode usesa pressurized fluid stream to position thesample plug aspirated from the sample vialinto the injection loop. In the needle-overfillpartial loop mode, the syringe draws anexcess of sample into the needle and throughthe valve while the loop remains in-line withthe pump. The valve is switched to bring theloop off-line and the syringe then meters theappropriate volume of sample into the loop.The valve is then switched back again tocomplete the injection.

To evaluate injection precision, samples ofphenol and toluene at concentrations rang-ing from 0.2 to 9.0 mg/mL, depending onthe injection volume, were prepared in50/50 v/v acetonitrile/water. A set of sixinjections was performed at each of the fourinjection volumes: 0.5 mL, 1 mL, 2 mL, and5 mL; on 5-mL, 10-mL, and 20-mL loops.Partial loop injections were evaluated in thepressure-assisted mode on all three loopsizes; needle-overfill and full loop modeswere also evaluated on the 5-mL loop. Iso-cratic elution (30/70/0.1 v/v/v acetoni-trile/water/TFA) at a flow rate of0.8 mL/min was used to elute the com-pounds off a 2.1 3 50 mm, 1.7-mmACQUITY UPLC BEH C18 column. Thecolumn temperature was 35 °C, and detec-tion was performed at 270 nm (phenol) and261 nm (toluene) with 10 pts/s data collec-tion rate. The peak areas were measured foreach injection and the relative standard devi-ations were determined. The results are dis-played in Table II.

The best precision was obtained with afull loop 5-mL injection, where six injectionsproduced 0.3% RSD for both analytes. Thiswas expected, as overfilling the loop presentsless opportunity for variation than transfer-ring and positioning a sample plug in theloop for partial loop injections. The partialloop injections produced varied results. Allof the 0.5-mL injections yielded poor preci-sion (. 3% RSD in the pressure-assistedmode). The precision of partial loop injec-tions from 1–5 mL on the 10-mL and 20-mLloops ranged from 0.4-1.6% RSD. Valueson the higher end of this range may be unac-

ceptably high for typical system suitabilityrequirements of most validated assay meth-ods. One explanation for the poor precisionobserved is that in partial loop injections inthe pressure-assisted mode the sample plugcan potentially experience significant dilu-tion, which may result in some of the sam-ple diffusing out of the sample loop and notbeing injected onto the column. For this rea-son, Waters does not recommend injectingmore than 50% of the loop volume usingthe pressure-assisted mode. One wouldexpect the partial loop injection precision toworsen in a smaller loop as accurately posi-tioning the complete sample plug in theloop becomes more difficult. The needle-overfill mode may therefore be a moreattractive technique for use with the smallerloops. The sample experiences less dilutionin this mode and injections up to 75% of theloop volume may be performed. Moderateimprovements in precision were obtainedwith this mode compared to the pressure-assisted mode on the 5-mL loop.

As one can see, there are multiple factorsto consider that make the injection processnot a trivial matter. Users will have to keepthese in mind and determine for themselvesthe optimal scenario for the given method,which should include measuring the preci-sion for a particular injection volume, loopsize, and mode of injection. For work requir-ing high precision, it may be more appropri-ate to exclusively use full loop injections.Waters offers 2, 5, 10, and 20 mL loop sizes;enough choices to make that a viable option.

The gradient reproducibility of theACQUITY was also assessed. The ability of

The high resolutionobtained inextremely shortanalysis times makesUPLC a veryattractive tool forthe pharmaceuticaldevelopmentlaboratory.


the instrument to repeatedly produce anaccurate gradient is necessary to ensurereproducible retention times. The sampleanalyzed was a Novartis development prod-uct whose method was successfully trans-ferred to the ACQUITY UPLC system,described in detail below. This sample con-tained 12 components, each at a level of 1%of the target active concentration. A 10 mMdihydrogen phosphate buffer adjusted to pH 3 with phosphoric acid was used formobile phase A, and acetonitrile as mobilephase B. The gradient was 5–40% B in 2 min, and 40-60% B in 0.5 min, followed

by re-equilibration at starting conditions for0.5 min, for a total run time of 3.5 min.The flow rate was 0.8 mL/min, whichrequired a peak run pressure of 13,000 psi,and the column used was a 2.1 3 50 mmACQUITY UPLC BEH C18, held at 35 °C.A total of 200 2-mL injections on a 5-mLloop were performed in sequence over aperiod of more than 12 h, equal to . 5000column volumes of mobile phase pumpedthrough the column. Displayed in Table IIIare the retention time precision results for all12 peaks. The retention time repeatabilityfor 200 injections was <0.25 %RSD for all

12 peaks, indicating that no drift in the gra-dient occurred throughout the analysis.

Method TransfersFinally, evaluation of fundamental chro-matographic parameters can only go so far indetermining the usefulness of the technique.One must eventually test the instrument inthe manner in which it is to be used, namely,running test methods with actual samples.Figure 3a shows the analysis of a Novartisdevelopment combination product contain-ing three active pharmaceutical ingredientsand nine byproducts and degradation prod-ucts spiked at a level of 1% of the active con-centration. This method was originallydeveloped on a Waters® Alliance® Separa-tions Module HPLC using an XTerra™RP18, 3.5 µm, 4.6 3 100 mm column, andachieves the separation with a run time of 24min. To transfer this method to theACQUITY UPLC using a 2.1 3 50 mmACQUITY column, the flow rate and injec-tion volume were scaled in proportion to thedifference in column dimensions in order toobtain the same linear velocity. The originalgradient was simplified and adjusted, andthe flow rate was increased to obtain thefastest separation possible that still met theresolution requirements of 2.0 betweenpeaks 3 and 4. This utilized a flow rate of 1 mL/min at about 14,000 psi. The chro-matogram obtained with this new method isshown in Figure 3b. This sample is a forceddegradation of the three active componentsand is therefore of a different concentrationthan the sample used to obtain the chro-matogram in Figure 3a, which explains thelower signal-to-noise ratio observed. Thebenefits of the ACQUITY for fast LCmethod development is strikingly apparentfrom the run time of the new method devel-oped on the ACQUITY. At 2.8 min, agreater than 8-fold reduction in analysistime is achieved. Most importantly, ade-quate resolution (Rs 5 2.7) between the crit-ical peak pair (peaks 3 and 4) is maintained.

Due to potential selectivity differencesbetween columns, an eight-fold reduction inanalysis time may be too much to expect forall method transfers. It has been our experi-ence, however, that a five-fold improvementis possible for most reversed-phase LC meth-ods. This will not only enable rapid analysistimes, but will speed up the process of devel-oping methods. Provided the selectivity ofthe various ACQUITY columns can sepa-rate the sample components, transferringmethods already developed on conventional

(a)

(b)

Time (min)

Ab

sorb

ance

(A

U)

Ab

sorb

ance

(A

U)

0.014

0.012

0.010

0.008

0.006

0.004

0.002

0.000

-0.002

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4

0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000

-0.001

1

12

11

10

98

76

543

2

1

12

1110

98

765

4

3

2

RS = 2.7

2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0

25 mL of a 1% impurity spiked sample

2 mL of a forced degradation sample

Figure 3: Method transfer of a Novartis development combination product from a Waters 2690with an XTerra column (a) to the ACQUITY UPLC (b). The signal-to-noise difference observed isbecause the two samples are of different concentrations (see text for details). Conditions: (a)Waters XTerra RP18, 3.5 mm, 4.6 3 100 mm column, 35 °C, 25 mL injection volume, 1.3 mL/min;mobile phase: A – 92:8 pH 3 sodium phosphate buffer: acetonitrile; B – 50:50 pH 3 sodium phos-phate buffer:acetonitrile. (b) ACQUITY 1.7 mm, 2.1 3 50 mm column, 35 °C, 2 mL injection volumeon a 5-mL loop,1.0 mL/min; mobile phase: A – pH 3 sodium phosphate buffer; B – acetonitrile.


instruments and columns to the ACQUITYis straightforward and in most cases could beaccomplished within a few hours.

The high efficiency afforded by theACQUITY columns is also expected toincrease sensitivity—an analyte will be moreconcentrated in a narrower peak and thusgive rise to a higher signal. Limit of quanti-tation solutions of one of the active com-pounds from the previous method were pre-pared at 0.05 % and 0.1 % of its targetconcentration. The injection volume wasscaled in proportion to the column dimen-sions. These gave rise to signal-to-noiseratios of 15 and 22, respectively, using theoriginal 24 min method with a conventionalinstrument and column. When analyzedwith the new ACQUITY method, signal-to-noise ratios of 26 and 38 were obtained—animprovement in sensitivity by a factor ofabout 1.7. This is not an entirely direct com-parison, since the instruments, detectors,and data collection rates used were all differ-

ent, as was the gradient. Regardless of all thevariables that can affect peak height and sig-nal-to-noise, a new method was developedthat produced greater sensitivity than theformer method.

ConclusionThe high resolution obtained in extremelyshort analysis times makes the UPLC a veryattractive tool for the pharmaceutical devel-opment laboratory. The increasing numberof column chemistries available in the1.5–2 mm size range make the ACQUITY aflexible method development tool applicableto a wide range of samples. With properconsideration of factors affecting band-broadening, injection and gradient preci-

sion, as well as other chromatographic char-acteristics, the analyst can achieve fast, effi-cient, and reproducible methods. The resultsare compelling that 1.5 to 2 mm sizedporous stationary phase particles used withhigh-performance instrumentation such asthe ACQUITY UPLC will become theoption of choice for the development of fastLC methods in pharmaceutical developmentin the near future.

AcknowledgmentsThe authors would like to thank Liz Robert-son, Patricia McConville, and Vladimir Bin-shtock of the Waters Corporation for all oftheir discussions, assistance, and cooperationduring the beta evaluation of the ACQUITYUPLC. They would also like to thank theircolleagues at Novartis, specifically MinYang, Guy Yowell, and Yuri Kazakevich fortheir support and discussions concerningthis work.

References(1) F. Gerber, M. Krummen, H. Potgeter, A. Roth,

C. Siffrin, and C. Spoendlin, J. Chromatogr., A

1036, 127-133 (2004).

(2) http://www.acdlabs.com/products/chrom_lab

(3) http://www.rheodyne.com/products/chro-

matography/drylab

(4) http://www.iristechnologies.net/Chrom

Sword/ChromSwordAuto.htm.

(5) N. Tanaka, H. Kobayashi, K. Nakanishi, H.

Minakuchi, and N. Ishizuka, Anal. Chem. 73,

420A–429A (2001).

(6) N. Wu, J. Dempsey, P.M. Yehl, A. Dovletoglu,

D. Ellison, and J. Wyvratt, Anal. Chim. Acta

523, 149–156 (2004).

(7) C. Zhu, D.M. Goodall, and S.A.C. Wren,

LCGC 23(1), 54–72 (2005).

(8) T. Greibrokk and T. Andersen, J. Chromatogr,. A

1000, 743–755 (2003).

(9) A.D. Jerkovich, J.S. Mellors, and J.W. Jorgen-

son, LCGC 21(7), 600–610 (2003).

(10) M. Swartz, LCGC 23(1), 46–53 (2005).

(11) A.D. Broske, et al., Agilent Technologies appli-

cation note 5988-9251EN (2004).

(12) J.E. MacNair, K.C. Lewis, and J.W. Jorgenson,

Anal. Chem. 69, 983–989 (1997).

(13) J.E. MacNair, K.D. Patel, and J.W. Jorgenson,

Anal. Chem. 71, 700–708 (1999).

(14) L.T. Tolley, J.W. Jorgenson, and M.A. Moseley,

Anal. Chem. 73, 2985–2991 (2001).

(15) J.S. Mellors and J.W. Jorgenson, Anal. Chem.

76, 5441–5450 (2004).

(16) K.D. Patel, A.D. Jerkovich, J.C. Link, and J.W.

Jorgenson, Anal. Chem. 76, 5777–5786 (2004).

(17) J.A. Lippert, B. Xin, N. Wu, and M.L. Lee, J.

Microcolumn Sep. 11, 631–643 (1997).

(18) N. Wu, D.C. Collins, J.A. Lippert, Y. Xiang,

and M.L. Lee, J. Microcolumn Sep. 12, 462–469

(2000).

(19) N. Wu, J.A. Lippert, and M.L. Lee, J. Chro-

matogr., A 911, 1–12 (2001).

(20) L.R. Snyder, J.J. Kirkland, and J.L. Glajch, Prac-

tical HPLC Method Development, 2nd ed. (John

Wiley & Sons, Inc., Hoboken, New Jersey)

1997.

(21) U.D. Neue, HPLC Columns (Wiley-VCH,

Hoboken, New Jersey, 1997). n

ACQUITY UPLC willbecome the option ofchoice for thedevelopment of fastLC methods inpharmaceuticaldevelopment in thenear future.


Ultra Performance Liquid ChromatographyCoupled to Orthogonal Quadrupole TOF–MS(MS)for Metabolite Identification

Iain Beattie, Karine Joncour,and Kim LawsonAstraZeneca R&D Charnwood,Physical & Metabolic Science,Loughborough, U.K., [email protected]

iquid chromatography coupledwith mass spectrometry(LC–MS) and tandem mass spec-

trometry (LC–MS-MS) is well established asthe main analytical technique capable ofproviding the level of information andthroughput required (1) in the pharmaceuti-cal industry. LC–MS for the identificationof drug metabolites was first used in the mid1980s (2,3) with the introduction of thethermospray source which allowed for therelatively easy interfacing of reversed-phaseLC solvent systems to mass spectrometers(4). In the late 1980s, the introduction ofthe atmospheric pressure ionization interfacetook this approach a step further in terms ofperformance and ease of use. Through the1990s, the main advances were instrumentperformance and capability with triplequadrupoles allowing structural informationto be easily obtained to assist in the identifi-cation of sites of metabolism. Ion trap massspectrometers allowed additional informa-tion to be generated by being able to carryout MSn experiments which could give addi-tional information to that obtained from thetriple quadrupole (5). The subsequent devel-opment of the hybrid quadrupole-orthogo-nal time-of-flight (TOF) mass spectrometertook the identification of unknown drugmetabolites into a new era by providingaccurate mass information for both themolecular ion and fragments thereby givingincreased confidence in structural assign-ments. This approach has also been used byEckers and colleagues for identifying trace

impurities in drug formulations (6).For discovery support where large num-

bers of compounds need to be screened fortheir metabolic fate, the initial driving forcewas high throughput, but this usually com-promised both chromatographic resolutionand sensitivity. With the metabolic com-plexity of some samples and the need formore sensitivity as levels being incubated invitro or dosed in vivo decreased, there hasbeen a trend toward using narrower borecolumns to gain more sensitivity and betterresolution. However this had the disadvan-tage of increasing analysis run times. Theintroduction of the Waters CapLC™ sys-tem at the end of the 1990s was ideal for thistype of work making the use of 1 mm i.d.columns relatively easy, routine, and reliable.However, average run times were increasedfrom less than 10 min to around 30 min,which could become a potential bottleneck.With the advent of the Waters ACQUITYUltra Performance LC (UPLC) system,there was the real possibility of using 1 mmi.d. columns at higher flow rates resulting inimproved chromatographic resolution,increased sensitivity and a return to runtimes of around 10 min or less (7,8).

Here, we will compare the UPLC systemwith the Waters CapLC set up in a column-switching mode which is the LC system cur-rently being used in our laboratory. Theresults will demonstrate that with UPLC,the analysis time for routine in vitro samplescan easily be reduced from 30 min to 10min, while for more complex matrices

In the drug discovery process the detection and identification of the mainmetabolic routes that a compound will undergo is crucial to identifying ametabolically stable drug candidate which is a requirement for mostdiscovery projects but not all. In the discovery phase it is important thatthe main metabolites are detected and identified as rapidly as possible toallow the result to feed back into the cycle and influence the syntheticchemistry to either block a potential metabolic route to reduce clearance oravoid a substructure that might result in a potentially reactive or activemetabolite.

L


improved separation can be achieved relativeto the CapLC.

ExperimentalChemicals: Methanol (HPLC grade) andammonium acetate were purchased from

Fisher Scientific, U.K. Leucine-enkephalinwas obtained from Sigma-Aldrich.

Liquid ChromatographyHPLC was carried out on a Waters CapLCliquid chromatograph (Waters Corporation,

Milford, Massachusetts), which was config-ured for a column-switching mode. TheHPLC trapping cartridge was a 10 3 1 mmHyPURITY C18, 3.5 mm (Thermo Elec-tron Corp.) while the analytical column wasa 150 3 1 mm HyPURITY C18, 3.5 mm

100

%

0

7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0

100

%

0

Six oxidized metabolitesPeak width: 13 s

Summed ion trace forother metabolites

Time (min)

100

%

0

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

100

%

0

Time (min)

Six oxidized metabolitesPeak width: 5.5 s

Summed ion trace forother metabolites

Figure 1: Ion traces obtained for metabolites on the CapLC.

Figure 2: Ion traces obtained for metabolites on the UPLC.

© 2005 Waters Corporation. Waters, ACQUITY Ultra Performance LC, ACQUITY UPLC and LCT Premier are trademarks of Waters Corporation.

UPLC™ DATA

HPLC DATAHPLC DATA

UPLC™ DATA

SEE ALL THAT YOU'VE BEEN MISSING You’re expected to produce more chromatographic information about your samples in less

time. Until now, this seemed impossible due to high throughput performance compromises

with conventional HPLC. What if you could simultaneously maximize your data quality and

improve your cycle times? You can. Better analyses with more information…all in less time.

It’s just the tip of the iceberg with the ACQUITY UPLC System.

MORE DATA = MORE INFORMATION Introduced in 2004, the Waters ACQUITY Ultra Performance LC™ System is based on a

revolutionary holistic instrument design concept, built from the ground up to re-invent liquid

chromatography as you know it. Comprising innovative column chemistries, hardware

and software, the ACQUITY UPLC is a first-of-its-kind LC system optimized to leverage the

potential of patented sub-2-micron particle technology. This concept enables superior

chromatographic separations that are up to nine times faster than today's HPLC systems,

with 2-fold better resolution and 3-fold better sensitivity. This means reliable, high quality

sample information in less time for truly higher productivity, without compromise.

MORE KNOWLEDGE = BETTER DECISION MAKINGMaking more informed decisions in less time – a universal benefit in today’s laboratory.

Whether you’re a methods developer, researcher, mass spectrometrist or a chromatographer,

you seek a productivity edge. Now you can get more information in a single short run than

you’ve ever seen with your current HPLC methods. Smaller particles operate more effectively

under high linear velocities, yielding greater efficiency and allowing you to benefit

from simultaneous improvements in speed, sensitivity and resolution. The result is

unprecedented improvements in the quality of your data with the added benefit of

drastically reduced analysis times.

A growing number of scientists around the world are already experiencing the advantages of

gaining more quality information from their most challenging separations. See more, know

more – the future of liquid chromatography is here today with the Waters ACQUITY UPLC

System. To learn more, visit www.waters.com/acquity

The Waters ACQUITY UPLC System adds a new dimensionto complex mixture analysis. To illustrate, an LC/MS analysisof a biofluid is displayed in a 3-D plot format. The extra resolvingpower provided by a 1.7 µm UPLC separation (bottom) is compared to a conventional 3.5 µm separation (top).UPLC clearly provides superior chromatographic performance,yielding considerably more information about the sample.

- Courtesy of Professor Ian Wilson, AstraZeneca Pharmaceuticals

THERE’S MORE INFORMATION OUT THERE.

WATERS®

ACQUITY UPLC™ SYSTEM IS

THE ONLY ONE TO HELP YOU FIND IT.

THERE’S MORE INFORMATION OUT THERE.

WATERS®

ACQUITY UPLC™ SYSTEM IS

THE ONLY ONE TO HELP YOU FIND IT.

HPLCHPLC

UPLC™UPLC™

“UPLC revealed peaks in the samplesthat we simply didn’t know werethere by conventional LC.”

-Professor Ian Wilson, AstraZeneca Pharmaceuticals


the analytical column with a flow rate of 40ml/min running a gradient from 5–95%methanol, using 0.025% ammoniumacetate in water as the aqueous phase with atotal run time of 27 min.

UPLC was carried out on a Waters®ACQUITY UPLC™ system, without col-umn-switching, using a 100 3 1 mm WatersACQUITY UPLC BEH C18 1.7 mm col-umn with a flow rate of 110 ml/min. The

HPLC column (Thermo Electron Corp.).The sample was loaded at 20 ml/min in0.25% ammonium acetate in water with 2%methanol for 2 min. The valve was thenswitched and the sample back flushed onto

100

%

0

100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950

100

%

0

m/z

UPLC

CapLC

Figure 3: Mass spectra of oxidized metabolite from UPLC and CapLC.

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0

Time (min)

Figure 4: Selected ion traces for rat urinary metabolites from the UPLC.


same solvents were used for both theUPLC™ and the CapLC. The in vitro sam-ples were analysed using a total run time of10 min, while for the in vivo samples it was20 min.

The HPLC column was also run on the

UPLC system using the same solvents andgradient as the CapLC but with a higherflow rate of 55 ml/min, but without column-switching.

Radio-flow DetectionThe detector was a Lablogic b-RAM(Sheffield, U.K.) fitted with a 50 ml liquidscintillant flow cell. The scintillant flow ratewas 0.5 ml/min for both UPLC and CapLCruns.

100

%

0

5 6 7 8 9 10 11 12 13 14 15 16 17 18Time (min)

100

%

0

100

%

0

100

%

0

100%

0

100

%0

Figure 5: Selected ion traces for rat urinary metabolites from the CapLC, where metabolite at m/z 515 was not detected.

100

%

0

100

%

0

100

%

0

100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560

m/z

(a)

(c)

(b)

Figure 6: Nonbackground subtracted spectra obtained from rat urine: spectrum A is the metabolite only found from the UPLC data, spectrum B isa metabolite from the CapLC data with the corresponding spectrum from the UPLC data in C.


Mass SpectrometryAll the MS was carried out on a WatersMicromass® Q-Tof II (Waters, Manchester,U.K.) in either positive or negative ionmode. The analyses were carried out usingleucine-enkephalin as the lock-mass via theLockSpray interface. All data were collectedin centroid mode. The TOF data were col-lected between m/z 100 and 950 with anacquisition rate of 0.4 s/spectrum for mostof the data with a rate of 0.9 s/spectrum forthe rest.

SamplesThe samples used were examples taken fromthe routine work flow and were either ratliver hepatocyte incubations or rat urineobtained from a bile duct cannulation study.

Results and DiscussionThe two main advantages of UPLC over theCapLC is the increased resolution thatenables faster analysis times. Figure 1 showsthe type of separation from an in vitro sam-ple incubated with rat hepatocytes. This par-ticular compound generated a significantnumber of metabolites including six oxi-dised species which were well separated onthe CapLC. The UPLC traces in Figure 2show comparable resolution although theresolution between the six oxidised specieson the UPLC could be improved with fasterscan rates as the peaks are quite narrow.However the data from the UPLC were gen-erated in 10 min instead of 30 min. In addi-tion to the speed at which the data can beobtained, the quality of the data is alsoimproved. To compare the quality of thespectra from the two systems, Figure 3shows the spectra for one of the oxidizedmetabolites taken from the top of the peak.It is clear that the quality of the UPLC spec-trum is better than that from the CapLCwith much improved signal-to-noise.

Although screening in vitro samples is asignificant part of the work process, as proj-ects progress or there are particular issues toaddress, there is a need to look at in vivosamples to see how the metabolic profilecompares with the in vitro results. Thesematrices are most commonly urine or bileand to ensure that there are no losses of anymetabolites they are subjected to no samplework up other than centrifugation for urineand dilution with water followed by cen-trifugation for bile. However, the level of thebackground endogenous material found inthose samples is much higher than thatfound in in vitro preparations. This is where

the increased resolution and peak capacity ofthe UPLC system should have an impact byproviding greater separation of compound

related material from endogenous with lesslikelihood of interference or possible sup-pression. For this type of sample good sepa-

350.0

300.0

250.0

200.0

150.0

100.0

50.0

0.0

320.0

280.0

240.0

200.0

160.0

120.0

80.0

40.0

0.0

cpm

cpm

0.00 5.00 10.00 15.00 20.00

Time (min)

0.00 5.00 10.00 15.00 20.00

100

%

0

4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0

100

%0

Time (min)

Figure 7: Summed selected ion traces from a rat urine sample from the (top) CapLC and the (bot-tom) UPLC.

Figure 8: Radio traces for a rat urine sample obtained from the (top) CapLC and the (bottom)UPLC.


ration and sensitivity is more important thanspeed of analysis.

The selected ion traces in Figure 4 are fora number of metabolites found in rat urinefrom the UPLC, while the correspondingresults obtained from the CapLC are shownin Figure 5. From these traces there are twoobservations. The first is that the metabolitewith m/z 515 was not detected at all in theCapLC data, possibly as a result of ion sup-pression. The second is that there is an

obtained from the UPLC compared to theCapLC. The spectra are single spectra withno background subtraction with spectrum Bfrom the CapLC and spectrum C compar-ing the same metabolite from the UPLC.From this, it is quite clear that the UPLCgives a much higher quality spectrum withthe base peak being the molecular ion andless endogenous interference while in theCapLC spectrum the metabolite is less than30% of the base peak which is an endoge-nous interference. Spectrum A is themetabolite that could only be detected fromthe UPLC data and again shows the qualityof the spectrum with the isotope patterncomfirming it is compound related.

To determine the fate of drug candidatesin vivo, it is best to work with radio-labelledmaterials. This approach can provide a moredefinitive quantitative measure but also indi-cate where to look in the mass spectral datafor metabolites. To do this requires runningthe sample through a radio-flow detectorand matching the radio traces obtained withthe mass spectral data. The radio detectorcan contribute to some peak broadening asthe peaks are being mixed with scintillant soincreased resolution and sharper peaks canimprove the quality of the radio trace. Figure7 shows the summed selected ion traces for arat urine sample from the CapLC and theUPLC both with a 20 min run time. Typicalpeak widths from the CapLC are 12–15 swhile for the UPLC this was reduced to 6-10s. By comparing the radio traces in Figure 8the peak widths are 30 s for the CapLC and20 s for the UPLC data. In addition, peak 4in the MS traces did not produce a peak inthe radio trace from the CapLC data butdoes in the UPLC trace.

As a final comparison, a rat hepatocyteincubation was run on the CapLC, theACQUITY UPLC system using theACQUITY column and also on the UPLCusing the HPLC C18 column to determinehow an ordinary column would perform onthe UPLC system. The resultant traces areshown in Figures 9 (CapLC), 10 (UPLCwith ACQUITY column), and 11 (UPLCwith HPLC C18). The peak widths for thevarious traces were around 6–7 s for theUPLC with ACQUITY column and 12–15 s for the HPLC column on both theCapLC and UPLC. The best results wereobtained on the ACQUITY column with allpeaks being resolved and the best spectralquality, as shown in Figure 12, for the weak-est metabolite with m/z 502. Increased sensi-tivity when using the ACQUITY system

endogenous component with m/z 419 elut-ing at 15.32 min close to the metabolite at14.95 min. This endogenous component ispresent in the UPLC trace but elutes aroundat 12.73 min, well separated from the peakof interest at 10.48 min. One issue with invivo samples is the level of the backgroundthat the endogenous components generatemaking it more difficult to pick out com-pound related peaks. Figure 6 demonstratesthe quality of the spectra that can be

100

%

011 12 13 14 15 16 17 18 19 20

Time (min)

100

%

0

100

%

0

100

%

0

100

%

03.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5

Time (min)

100

%

0

100

%

0

100

%

0

Figure 9: Rat hepatocyte incubation run on the CapLC.

Figure 10: Rat hepatocyte incubation run on the UPLC with the ACQUITY column.

UPLC–MS does offer significant gains overthe CapLC–MS for metabolite identifica-tion. For screening in vitro samples it is pos-sible to reduce analysis times by two thirdswhile still retaining chromatographic resolu-tion and sensitivity with an improvement inspectral quality. For the analysis of morecomplex in vivo samples it does offer bigimprovements in separation and again thegain in spectral quality is even more crucialmaking it much easier to detect compoundrelated peaks in the first place. From the datashown it also allowed the identification of ametabolite that was not even detected whenusing the CapLC. The sharper peaks alsogave added benefit when generating radiochromatograms to compare with the massspectral data. As a final point, it is also a veryeffective and capable HPLC and shouldallow for the replacement of other systemswith this giving the choice of using conven-tional HPLC columns up to their recom-mended pressure limits, typically 4000 psi,as well as the high pressure capable columns.The very low dead volume in the systemmakes this an excellent HPLC system withthe significant advantage over conventionalHPLC systems of having the high pressurecapability.

To push the UPLC capability further itwould be necessary to have a mass spectrom-eter that is capable of scanning fast enoughto be sure of getting good quality data.However we have clearly demonstrated thatusing the Q-Tof II it does give significantadvantages in chromatographic resolution,sensitivity and a big reduction in analysistime.

References(1) M.S. Lee, LC/MS Applications in Drug Develop-

ment (John Wiley, Hoboken, New Jersey), 2002.

(2) T.J.A. Blake, J. Chromatog. 394, 171–181,

1987.

(3) I.G. Beattie and T.J.A. Blake, J. Chromatog. 474,

123–138, 1989.

(4) C.R. Blakley, J.J. Carmody, and M.L. Vestal, J.

Am. Chem Soc. 102, 5931, 1983.

(5) G.J. Dear, J. Ayrton, R. Plumb, and I.J. Fraser,

Rapid Commun. Mass Spectrom. 13, 456, 1999.

(6) E. Eckers, N. Haskins, and J. Langridge, Rapid

Commun. Mass Spectrom. 11, 1916, 1997.

(7) R. Plumb, J. Castro-Perez, J. Granger, I. Beattie,

K. Joncour, and A. Wright, Rapid Commun.

Mass Spectrom. 18, 2331, 2004.

(8) R. Plumb, J. Castro-Perez, J. Granger, I. Beattie,

K. Joncour, and A. Wright, Rapid Commun.

Mass Spectrom. 19, 1, 2005. n

UPLC system did perform comparably tothe CapLC when used as an ordinaryHPLC.

ConclusionFrom the data shown it is clear that

with the ACQUITY column was alsoobserved where a three-fold gain wasachieved compared to the CapLC system.Additionally this gave the fastest analysistime of 10 min compared to 27 min for theHPLC column. It was also observed that the

11 12 13 14 15 16 17 18 19 20

Time (min)

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100

%

0

100 140 180 220 260 300 340 380 420 460 500 540 580

m/z

UPLC ACQUITY column

UPLC HyPURITY

CapLC HyPURITY

Figure 11: Rat hepatocyte incubation run on the UPLC with the HyPURITY C18 column.

Figure 12: Comparison of nonbackground subtracted spectra for peak with m/z 502 from thethree systems.

www.chromatographyonline.com30 SEPARATION SCIENCE REDEFINED MAY 2005


Assay Transfer from HPLC to UPLC forHigher Analysis Throughput

Ying Yang and Craig C.HodgesAlexza Molecular DeliveryCorporation, Palo Alto,California 94303, e-mail:[email protected].

ncreasing demand for greater phar-maceutical analysis throughputprompted the testing of the Waters

ACQUITY Ultra Performance LC(UPLC™). This system claims to providefaster analyses through the use of a novelseparation material of very fine particle size(1.7 mm) and unique core chemistry (1–5).

To effect fast separations on this material,the column hardware and instrument havesignificant design modifications from typicalHPLC. The UPLC operates at higher pres-sures (up to 15,000 psi.), injects samplesinto a smaller system dwell volume, and cap-tures detector signals at high data rates forfast eluting peaks. A new needle design hasbeen claimed to substantially reduce carry-over which can aid in the lowering of limitsof quantitation (LOQ).

In this work, an HPLC method for qual-ity control (QC) was optimized for UPLC.Strategies to reduce total runtime, lower costper assay, and promote instrument uptimewere considered.

Method DevelopmentThe original 10-min HPLC QC assay wasdeveloped to quantify the content of a hete-rocyclic drug (Cpd A) in organic solventextracts. An internal standard (IS) was usedto compensate for sample preparation lossesand a terminal washing gradient was neces-sary to remove late eluting interferences.

Initial transfer of the HPLC assay toUPLC was accomplished by simply applyinga scaling factor to the mobile phase flow rateand the sample injection volume. This scal-ing factor was derived from the ratio of thecolumn cross sectional areas in order toretain the mobile phase linear velocity.

Chromatograms from this UPLC methodhad very narrow peaks, and the excessive res-olution indicated opportunity for methodimprovement. The mobile phase flow ratewas increased until limited by column back-pressure. However, subsequent column life-time studies indicated that reducing totalrun time by increasing organic solvent con-tent was more economical. A dramaticdecrease in solvent consumption was alsoobtained. Chromatograms in Figure 1 com-pare the original HPLC method to those ofthe initial scaling and the final UPLC condi-tions. Parameters of the HPLC and finalUPLC methods are listed in Table I.

Method Optimization Guidelinesand ObservationsDuring the course of optimizing the UPLCmethod, considerations to expedite futuremethod transfers were developed, and thefollowing recommendations were made:● Increase elution solvent strength to

reduce run times taking advantage of thehigh resolution potential of UPLCcolumns (see Table II).

● Increase mobile phase flow rate secondar-ily to solvent strength in order to pro-mote longer column lifetimes. Whilehigh mobile phase linear velocities withgood resolution are possible (Figure 2),as with any column, routine operation at80% maximum rated pressure led toshortened lifetimes. In our experience,UPLC operation around 8000 psi or lessprovided comparable or lower columncost per assay than HPLC. Maintaininglow flows as much as possible alsoreduces solvent and waste disposal costs,although these are already an order of

A typical HPLC assay was transferred and optimized for a Waters ACQUITYUPLC™ system to achieve both higher sample analysis throughput andbetter assay sensitivity. Strategies to expedite future method transferswere compiled. Analysis of operation costs and sample throughput foundUPLC cost advantageous over HPLC.

I


magnitude less than HPLC.● Reduce column re-equilibration times by

taking advantage of the low system dwellvolume. Programmed changes in themobile phase take time to reach the col-umn. The small UPLC dwell volume

(measured as 110 mL, 15% of that of theHPLC) allowed in part the abbreviationof the original assay. Column re-equili-bration was accomplished during nextsample loading in the UPLC, furtherincreasing throughput.

● Reduce injection volumes appropriatelyfor the column diameter to achieve goodpeak shapes. Peak splitting can occurwhen too large of a strong sample solventbolus overwhelms the packing at the col-umn head. While this assay method tol-erated 5 mL injections, volumes of 1–3mL are more typical starting points inour experience. Note that smaller injec-tion volumes may be compensated byenhanced peak height from use of thehigh resolution columns and by the lowcarryover from the UPLC injector (meas-ured as 10% of the HPLC carryover forthis analyte) to achieve an equivalent oreven lower LOQ). An alternative tosmaller injection volumes might be tolower sample solvent strength to accom-plish sample focusing on the head of thecolumn.

● Utilize partial loop-fill injections in pref-erence to full loop-fill. Partial loop-fillprecision was good even at volumes upto 80% of the loop total volume (Figure3). Typical laboratory practice is to limitsample volume injections to roughly50% of the total loop volume. TheUPLC injection system, which utilizesair-gap sandwiching of the sample, allowsbetter utilization of the sample loop and

Table I: Original HPLC versus optimized UPLC assay parameters

HPLC Assay UPLC Assay

Column XTerra C18, 50 3 4.6 mm, ACQUITY UPLC BEH C18, , 50 34 mm particles 2.1mm, 1.7 mm particles

Flow Rate 3.0 mL/min 0.6 mL/minNeedle Wash Methanol Strong Needle Wash: 200 mL

Methanol; Weak Needle Wash: 600 mL ACN:H2O 10:90

Injection Volume 20 mL 3 mL partial loop fill or 5 mL fullloop fill with automatic overfill

Gradient T0 (25:75), T6.5 (25:75), T0 (36:64), T1.1 (95:05), (time in min) T7.5(95:5), T9 (25:75), T10 T1.3 (36:64)(ACN:H2O) (25:75)Flow Rate 3.0 mL/min 0.6 mL/minTotal Run Time 10 min 1.5 minTotal Solvent Consumption Acetontrile:10.5 mL Acetonitrile: 0.53 mL(including 0.5 min of delay Water: 21.0 mL Water: 0.66 mL time in between injections)Plate Count for Cpd A 2000 7500USP Resolution 3.2 3.4LOQ ;0.2 mg/mL 0.054 mg/mLCarry-over , 0.05% with needle wash 0.01%Delay Volume ;720 mL ;110 mL

Figure 1: Chromatograms (from top to bottom): original HPLC, initial scaling to UPLC showing peak shape improvement and possibility for furthermethod optimization, and final UPLC method. Order of peak elution: internal standard (IS) then Cpd A.

0.5

1.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

0.1

0.1

0.2

0.0

Ab

sorb

ance

(A

U)

0.0

Time (min)

Ab

sorb

ance

(A

U)

Ab

sorb

ance

(A

U)


Figure 2: A van Deemter plot derived from UPLC experimental data indicates that use of high flow rates is a plausible strategy to decreasing over-all runtime. This should be balanced with backpressure effects on overall column lifetime (see text).

higher injection precision, reducing theneed for use of the full loop-fill mode.From a practical point of view, full loopfill requires substantially greater samplemovement considering overfill functions.This likely increases subsequent needlewashing, which may impact samplethroughput and increase wear of thewashing hardware. Larger sample vol-ume transfers also increases exposure tosample particulates, lowering long-terminstrument reliability.

● If full loop-fill mode is utilized, perhapsfor very high precision requirements,ensure adequate loop overfilling. A sig-nificant laminar flow velocity differentialin the loading sample between its wallinterface and center is created in the verynarrow bore tubing of the UPLC injec-tor. Overfilling the sample loop by atleast four loop volumes was found neces-sary to fully displace wash solvent fromthe 5 mL injector loop. For this instru-ment, the manufacturer has determinedand set as the default the optimum over-fill volume with typical sample solventsfor each sample loop size. Operators canspecify other overfill volumes for unusualsample compositions.

● Choose the proper composition and vol-ume of weak sample wash to obtain goodpeak shape. A portion of the weak sam-ple wash solvent will be co-injected withpartial-loop filled samples. The weak sol-vent wash should therefore mimic theinitial conditions mobile phase in solvent

8.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

Injection volume (mL)

Peak

are

a (3

105 )

0 1 2 3 4 5 6

y = 182351x - 37630R2 = 0.9951

7

6

5

4

3

Linear velocity (mm/s)

HET

P (m

m)

0 1 2 3 4 5 6

Figure 3: Peak area data generated by partial loop-fill mode of a 5 mL nominal (4.8 mL actual)sample loop in the UPLC injector. For standard loop injectors, the deviation from linear injectionvolume, as seen above in the 5-mL injection, occurs at much lower loop utilization so the generalrule is to only load 40–50% of the loop capacity.

Table II: Adjusting mobile phase parameters utilizing resolution potential of UPLC

HPLC Original UPLC Initial Scaling UPLC Final

Flow Rate (mL/min) 3.0 0.63 0.60% ACN in 25 25 36Mobile Phase

Plate Count for 2000 9100 7500Cpd AUSP Resolution 3.2 6.7 3.4Between A and IS


Figure 4: Linear correlation between concentration and peak area at lower concentrations from 0.054 to 1.30 mg/mL (R2 5 0.996 with 1/X2weighting).

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.00 0.20 0.40 0.60 0.80 1.00 1.20

Are

a (3

105 )

Concentration (mg/mL)

strength. Utilizing the weak wash solventas a sample diluent in the sample loopmay enhance sample focusing onto thecolumn. The volume of the weak washmust be sufficient to purge the formerstrong wash solvent from the loop.

Preliminary Method ValidationPreliminary assessment was made of the newassay and the instrument for linearity andlinear range, precision, accuracy, system suit-ability, and sample carry over.

Linearity and Lower Limit ofQuantification (LLOQ) With the potential greater sensitivity ofUPLC, the scope of the assay applicationwas broadened to address samples whichcould differ in concentration by 500-fold.The same UPLC separation method was cal-ibrated and found acceptably linear for twoassay ranges (Figures 4 and 5). With anLLOQ of 54 ng/ml, the low range UPLCassay allowed analyses more typicallyaddressed by liquid chromatography–massspectrometry. Notably, this particular UPLCsystem is configured with a photodiode arraydetector. Use of a wavelength-specific detec-tor could provide an even lower limit ofquantification.

Precision and AccuracyTriplicate injections were made at specifiedconcentrations to assess precision (repeata-bility) and accuracy. Precision was evaluated

Table IV: Evaluation of precision and accuracy for low range calibration

Theoretical Peak Area Precision1 Calculated Conc. Accuracy2

Conc. (mg/mL) (% RSD)/Result (mg/mL) (% Deviation)

0.645 62518 0.3 0.646 0.362173 0.642 20.462264 0.643 0.2

1.30 115988 0.1 1.308 0.7114863 1.294 20.4115826 1.306 0.5

5.18 428428 0.1 5.174 20.1428756 5.178 20.03429553 5.188 0.2

25.8 2094015 0.2 25.78 20.072088395 25.71 20.32097868 25.83 0.1

1 Acceptance criterion: , 5.0%, all passed.2 Acceptance criteria: 65.0%, all passed.

Table III: Evaluation of precision and accuracy for low range calibration

Theoretical Peak Area Precision1 Calculated Conc. Accuracy2

Conc. (mg/mL) (% RSD)/Result (mg/mL) (% Deviation)

0.054 11834 4.5 0.0484 210.312127 0.0519 23.812897 0.0611 13.2

0.325 35647 3.3 0.332 2.334344 0.316 22.536703 0.345 6.2

0.645 62518 0.3 0.652 1.262173 0.648 0.662264 0.649 0.8

1.3 115988 0.1 1.290 20.7114863 1.277 21.8115826 1.288 20.9

1 Acceptance criterion: , 5.0%, all passed.2 Acceptance criteria: 65.0%, except lowest concentration 615.0%, all passed.


Figure 5: Linear correlation between concentration and peak area at higher concentrations from 0.325 to 25.8 mg/mL (R2 5 0.999967 with 1/X2weighting).

0.0

0.5

1.0

1.5

2.0

2.5

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00

Are

a (3

106 )

Concentration (mg/mL)

by the peak area relative standard deviation(RSD). Accuracy was assessed by back-calcu-lation of the injection peak areas using thecalibration curve to give the calculated con-centration for each injection. These valueswere compared to the theoretical value andreported in terms of % deviation from thetheoretical value. The results for both lowand high range assays passed acceptance cri-teria (Tables III and IV).

System SuitabilityFive replicate injections were made to evalu-ate system suitability. The results passed allthe common USP acceptance criteria(Table V).

Injection–to-Injection SampleCarry-Over Contamination of a sample injection byresidues of the previous sample in the instru-ment (carry-over) can set the boundary foran assay’s LLOQ. Carry-over frequentlyleads to failure of tests for precision, accu-racy, and system suitability. However,depending on the protocol details of thesestudies, significant carry-over effects maynot be revealed. Direct measurement ofcarry-over was performed here to anticipateinaccuracies arising in potentially mixed setsof concentrated and dilute samples.

The UPLC instrument had design fea-tures to reduce sample carry-over: a novelneedle-in-needle injector design as well astwo separate injector wash solvents. In thisassay, 200 mL methanol were used as thefirst wash to remove the bulk of organicresidues, followed by 600 mL water:ACN(90:10) to displace the strong solvent and

bring the remaining sample loop, needle,and valve solutions to a composition com-patible with initial method conditions.

Carry-over was evaluated here by analyz-ing a solvent blank sample after each of thecalibration standards and measuring the areaof any peak appearing at the analyte reten-tion times. No interference peak wasdetected in the blanks run after the fivelower concentration standards. For blanksrun after injections of the highest concentra-tion standard, faint peaks slightly abovenoise were measured at 0.01% of the analytepeak in the previous injection. This wasacceptable for this assay, although carry-overmay have been reduced further by optimiz-ing the wash solvent parameters. In compar-ison, carry-over on the HPLC system was 5to 10 fold higher.

SummaryA QC HPLC assay to quantitate a hetero-cyclic pharmaceutical in organic solventextracts has been successfully transferred andoptimized for UPLC. Preliminary assess-ment indicates that the assay can be vali-

dated. Guidelines to expedite the develop-ment of future UPLC assays were compiled.The application of UPLC will be cost advan-tageous. While UPLC column expense peranalysis will be comparable to or slightly lessthan HPLC, solvent consumption and wastedisposal charges should decrease better thanan order of magnitude. Reduction of assaytime by five-fold dramatically improvesinstrument return on investment andreduces the total number of instrumentsneeded if only HPLC were employed.

References(1) A.D. Jerkovich, J.S. Mellors, and J.W. Jorgen-

son, LCGC 21(7), 660–611 (2003).

(2) N. Wu, J.A. Lippert, and M.L. Lee, J. Chromo-

togr. 911(1) (2001).

(3) K. K. Unger, D. Kumar, M. Grun, G. Buchel, S.

Ludtke, Th. Adam, K. Scumacher, and S.

Renker, J. Chromatogr., A 892(47) (2000).

(4) M. E. Swartz and B. Murphy, Lab Plus Int.,

18(6) (2004).

(5) M. E. Swartz and B. Murphy, Pharm. Formula-

tion Quality 6(5), p. 40 (2004). n

Table V: Evaluation of system suitability1

Injection Peak Area Plate Count USP Resolution

1 115988 7490 3.42 114864 7650 3.43 115827 7510 3.44 115896 7520 3.4 5 115104 7530 3.4

Average 115536 7540 3.4%RSD 0.4 0.8 0.4Acceptance %RSD , 2.0% Plate Count . ResolutionCriteria/Result /Pass 2000/Pass . 2.0/Pass

1 Replicate injections of 1.30 mg/mL standard.


The Evaluation and Application of UPLC forthe Rapid Analysis of Dose Formulations

Salane King, Peter J.Stoffolano, Eric Robinson,Thomas E. Eichhold, StevenH. Hoke II, Timothy R.Baker, Eloise C. Richardson,and Kenneth R. Wehmeyer Procter & GamblePharmaceuticals, Mason, Ohio

Address correspondence [email protected].

he pharmaceutical industry isunder intense pressure to increaseproductivity and put new drugs

onto the market in a shorter time period (1).Various approaches including high-through-put screening, combinatorial chemistry, pro-teomic/genomic target identification tech-niques, high-throughput in vitro screeningto determine physiochemical and absorp-tion/distribution/metabolism/eliminationproperties of compounds, rapid in vivopharmacokinetic screening, and the use ofbiomarkers and pharmacogenomics arebeing employed in drug discovery and devel-opment activities. A common theme inthese approaches is the need to provide high-quality data at a faster rate to drive decisionmaking processes.

Analytical chemists are challenged to findfaster ways of delivering quality data across arange of project-driven needs. A number ofapproaches are being employed to increaseseparation throughput including dual highperformance liquid chromatography(HPLC) column switching techniques (2,3),direct injection approaches for biologicalsamples (4,5), high-speed supercritical fluidchromatography (6), parallel (96-capillary)capillary electrophoresis (7,8), parallel (24-lane) HPLC (9,10), monolithic HPLCcolumns (11,12), and ultrahigh perform-ance chromatography (UPLC) (13–17).UPLC was pioneered in the late 1990s bythe Jorgenson (13–15) and Lee (16,17)groups and typically involves the use of 1–2mm particles at much higher pressures

(15,000–100,000 psi) than conventionalHPLC instruments (4000–6000 psi). Theuse of columns packed with 1-mm particlesin combination with high pressures allowsdramatic decreases in analysis time with lit-tle compromise in column performance.The relatively flat nature of the van Deemterplot for these small particles at higher linearvelocities accounts for the ability to operateat high flow rates without severe effects oncolumn efficiency (13,15). Recently, com-mercial instrumentation capable of operat-ing up to 15,000 psi combined withcolumns packed with 1.7 mm particles hasbecome available in the form of the Waters®ACQUITY UPLC™ system. This reportdescribes the basic performance of this com-mercial instrumentation, as well as pharma-ceutical applications for the rapid analysis ofdose formulations.

ExperimentalACQUITY UPLC PerformanceCharacterization: Isocratic Conditions: AnACQUITY UPLC system (Waters, Milford,Massachusetts) equipped with a UPLCBEH C18 column (2.1 3 50 mm, 1.7 mm)was evaluated under isocratic conditionsusing a water/acetonitrile/formic acid(65/35/0.1; v/v/v) mobile phase pumped atflow rates from 0.1 to 1.2 mL/min. The col-umn temperature, data sampling rate, filterconstant, injection volume, and detectionwavelength were 55 °C, 20 pts/s, fast, 5 mLand 233 nm, respectively. A mixture of NAP(10 mg/mL) and KR (100 mg/mL) in

The use of Ultra Performance Liquid Chromatography (UPLC™) with UVand MS detection was evaluated in several pharmaceutical applications.Initial studies characterized the commercial instrumentation performanceparameters for isocratic and gradient separations as a function of flow rateusing several commercially available drugs as model probe compounds.Parameters examined included reproducibility of retention time and peakarea, as well as theoretical plates, resolution, and column ruggedness. Theuse of high-speed separations for dose formulation strength analysis wasevaluated using two model drug compounds: mefenamic acid andchloramphenicol, in a dimethylacetamide/polyethyleneglycol-200 vehicle.Accuracy, precision, and resolution were examined under high-speedgradient conditions for the dose formulation analysis.

T


water with 0.1% formic acid (v/v), mobilephase B was acetonitrile with 0.1% formicacid (v/v) and the column temperature, datasampling rate, filter constant, detectionwavelength and injection volume were50 °C, 20 pts/s, fast, 277 nm and 5 mL,respectively. Replicate injections (n 5 5) of atest sample composed of 50 mg/mL each ofMFA and CA and 75 mg/mL NAP inwater/MeOH (50/50; v/v) was used toexamine performance under each gradientcondition.

Rapid dose formulation analysis: Amodel dose formulation containing bothMFA (1.410 mg/mL) and CA (0.856mg/mL) was prepared in DMAC/PEG200(10/90; v/v). Replicate (n 5 3) aliquots(1 mL) of the dose formulation was trans-ferred to separate 25-mL volumetric flasksand each was diluted to volume withwater/MeOH (50/50; v/v). A series of stan-dards containing both MFA and CA at the10, 25, 50, and 100 mg/mL levels were pre-pared by serial dilution of a combined1 mg/mL stock of each compound usingwater/MeOH (50/50; v/v) as the diluent.The diluted dose formulations and eachstandard were injected in replicate (n 5 5)on the system using the two gradient condi-tions described in Table I. The concentrationof MFA and CA in the dose formulation wasdetermined by interpolation from a standardcurve constructed using the first injection ofeach standard for the curve construction.

Results and DiscussionUPLC System Performance: Isocratic analysis:A standard solution containing KR andNAP as model drug compounds was exam-ined by UPLC under isocratic conditions todetermine the instrument performance as afunction of flow rates from 0.1 to 1.2 mL/min. Representative chromatogramsfor low, mid, and high flow rate conditionsare shown in Figure 1. Symmetrical peakshapes were obtained under all flow condi-tions and KR and NAP were well resolvedunder all conditions. The average values andreproducibility (%RSD) for the retentiontime, peak area, peak height, theoreticalplates (N), and resolution obtained for threereplicate injections under each flow condi-tion are shown in Tables II and III for KRand NAP, respectively. The RSD valuesobtained for retention time reproducibilitywere, in general, less than 0.1%. The peakarea and peak height reproducibility were, ingeneral, less than 1% RSD and most valuesranged between 0.1 and 0.4% RSD. Theo-

Gradient Time (min) Flow (mL/min) %A %B

Standard 0.00 0.50 75 252.00 0.50 0 1002.10 0.50 75 25

Fast 0.00 1.00 75 250.50 1.00 0 1000.51 1.00 75 25

Table I: Gradient conditions

Flow Rate tr, min Peak Area 3 104 Peak Ht 3 103 N

(mL/min) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD)

0.10 4.159 (0.06) 315 (0.28) 445 (0.38) 8878 (1.00)0.20 2.188 (0.03) 157 (0.23) 485 (0.71) 13257 (1.12)0.30 1.513 (0.04) 105 (0.04) 450 (0.23) 14136 (0.44)0.40 1.171 (0.05) 79.3 (0.03) 399 (0.39) 14196 (0.62)0.50 0.965 (0.00) 62.9 (0.22) 366 (0.26) 13261 (0.41)0.60 0.824 (0.07) 52.1 (0.17) 326 (0.28) 12186 (0.78)0.70 0.724 (0.00) 44.4 (0.13) 296 (0.13) 11465 (0.84)0.80 0.646 (0.09) 38.8 (0.23) 270 (0.34) 10801 (0.96)0.90 0.585 (0.20) 34.4 (0.24) 247 (0.20) 10091 (0.41)1.0 0.537 (0.11) 30.4 (0.65) 225 (0.50) 9447 (0.45)1.1 0.496 (0.00) 27.5 (0.47) 209 (0.47) 8994 (0.47)1.2 0.461 (0.13) 25.1 (0.54) 195 (0.57) 8374 (0.67)* tr 5 Retention Time, Peak Ht 5 peak height, and N 5 Theoretical Plates

Table II*: Ketorolac — Average (%RSD). Retention Time, Peak Area, Peak Height, and Theo-retical Plates (N) as a Function of Flow Rate (n 5 3 injections at each flow rate)

0.40

0.00

0.40

0.00

0.30

0.00

0.25

0.00

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00

0.20 mL/min

0.40 mL/min

0.80 mL/min

1.2 mL/min

KR NAP

Time (min)

Res

po

nse

(A

U)

Figure 1: Representative chromatograms obtained under isocratic conditions at various flowrates for a combined Ketorolac/Naproxen standard using UPLC.

MeOH was injected in triplicate at eachflow rate. The average and %RSD was cal-culated for the retention time, peak area,peak height, theoretical plates, and resolu-tion under each flow condition. Addition-ally, over 4000 injections were performedunder a high pressure conditions (1.25mL/min; 13,500 psi) to test the columnruggedness. Interspersed among the 4,000

injections were runs done at0.5 mL/min that were used to assess the col-umn performance.

Gradient Conditions: An ACQUITYUPLC system equipped with an UPLCBEH C18 column (2.1 3 50 mm, 1.7 mm)was evaluated under a standard and a fastgradient condition (see Table 1). For bothgradient conditions, mobile phase A was


retical plates decreased by 41 and 26% forKR and NAP, respectively, when operatingat the highest flow rate, 1.2 mL/min, relativeto the optimal values found at 0.3–0.4mL/min. Similarly, there was a 17%decrease in resolution when operating at thehighest flow rate relative to the optimal flowrate. This small decrease in resolution iscontrasted with the four-fold decrease inanalysis time.

The long-term performance of the UPLCBEH C18 column was determined by per-forming 4000 isocratic injections of the

dition) and fast (1.0 mL/min flow rate with0.5 min linear ramp from initial to finalmobile phase condition) gradient conditionswere examined using a probe sample con-taining CA (50 mg/mL), NAP (75 mg/mL),and MFA (50 mg/mL). Representative chro-matograms for the standard and fast gradi-ent conditions are shown in Figure 2. Allthree peaks are well resolved under both gra-dient conditions with the last peak, MFA,eluting at 1.54 and 0.56 min for the stan-dard and fast gradient, respectively. Usingthe fast gradient conditions, an injection-to-injection time of 1 min could be achieved.The average value and %RSD obtained forthe retention time, peak area, and resolutionfor each compound are shown in Table IV.The RSD values for the retention time of allthree compounds for the standard and fastgradient methods were less than 0.07% and0.30%, respectively. The peak area RSD val-ues for all compounds under standard andfast gradient conditions were less than0.70%. Resolution decreased from a value of13.3 under the standard conditions to 4.10under the fast gradient conditions but asseen in Fig 2 all peaks were well resolvedunder the fast gradient conditions.

Rapid Dose Formulation Analysis The analysis of CA and MFA in the

0.24

0.12

0.00

0.50

0.25

0.00

0.00 0.25 0.50 0.75 1.00 1.25 1.50

CANAP

MFA

MFANAP

CA

Time (min)

Res

po

nse

(A

U)

(a)

(b)

KR/NAP standard at high pressure(;13,500 psi, 1.25 mL/min). The columnperformance was periodically monitoredduring the injections by analyzing theKR/NAP standard at a lower flow rate(0.5 mL/min). The retention times, peakshapes, theoretical plates, and the resolutionwere found to remain constant throughoutthe 4000 injections (data not shown).

Gradient analysis: The performance of theUPLC instrument under standard(0.5 mL/min flow rate with 2 min linearramp from initial to final mobile phase con-

Figure 2: Representative chromatograms for a combined chloramphenicol/mefenamic acid obtained under: (a) fast and (b) standard gradient UPLCconditions.

Flow Rate tr, min Peak Area 3 104 Peak Ht 3 103 N Resolution

(mL/min) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD)

0.10 7.202 (0.08) 533 (0.30) 386 (0.99) 6320 (0.10) 11.2 (0.10)0.20 3.806 (0.03) 267 (0.21) 449 (0.21) 10287 (0.95) 14.3 (0.37)0.30 2.639 (0.04) 178 (0.13) 434 (0.11) 11650 (0.51) 15.2 (0.19)0.40 2.043 (0.05) 133 (0.17) 413 (0.38) 12185 (0.75) 15.6 (0.31)0.50 1.686 (0.03) 1064 (0.11) 400 (0.20) 11869 (1.13) 15.3 (0.44)0.60 1.440 (0.07) 87.9 (0.12) 366 (0.23) 11438 (0.78) 15.0 (0.33)0.70 1.262 (0.05) 75.3 (0.11) 346 (0.13) 11047 (0.38) 14.7 (0.35)0.80 1.124 (0.00) 65.9 (0.15) 325 (0.27) 10689 (0.42) 14.4 (0.34)0.90 1.015 (0.10) 58.1 (0.22) 305 (0.24) 10173 (0.49) 14.0 (0.16)1.0 0.928 (0.00) 51.1 (0.83) 283 (0.73) 9762 (0.27) 13.7 (0.08)1.1 0.853 (0.07) 46.2 (0.76) 267 (0.75) 9361 (0.46) 13.3 (0.10)1.2 0.791 (0.07) 42.3 (1.00) 254 (1.11) 8981 (0.47) 12.9 (0.13)* tr 5 Retention Time, Peak Ht 5 Peak Height and N 5 Theoretical Plates.

Table III*: Naproxen — Average (%RSD). Retention Time, Peak Area, Peak Height, Theoreti-cal Plates (N), and Resolution as function of Flow Rate (n 5 3 injections at each flow rate)


DMAC/PEG200 dose formulation wasdone under the standard and fast gradientconditions for three replicate dilutions of thedose formulation using replicate (n 5 5)injections for each dilution. The concentra-tions of CA and MFA in the diluted doseformulations were determined by interpola-tion from a linear regression standard curve(Table V). Very similar results were obtainedby the standard and fast gradient approachin terms of the found concentration,although the data for the fast analysis wasslightly more variable than the standard con-ditions but was still very acceptable for drugdiscovery efforts.

ConclusionsThe ACQUITY UPLC System was shownto provide accurate and reproducible resultsfor rapid isocratic and gradient analysis ofdrug molecules in dose formulations. In the

rapid isocratic mode, retention times, peakshapes, theoretical plates, and resolution fora multicomponent drug standard werefound to remain constant over 4000 injec-tions. Using fast gradient conditions, injec-tion-to-injection cycle times of 1 min couldbe obtained, while still maintaining accept-able within-specification results for repeata-bility for both peak retention time and area.The utility of the UPLC system for therapid, accurate and precise analysis of drugsin dosage formulations was demonstratedusing a model drug system. The UPLC sys-tem provided the same accuracy, precision,and ruggedness as a standard HPLC systembut allowed decreased analysis times result-ing in higher sample throughput.

References(1) D. Brown and G. Superti-Furga, Drug Discovery

Tech. 8, 1067–1077 (2003).

(2) G.A. Smith, C.M. Rawls, and R.L. Kunka,

Pharm. Res. 21, 1539–1544 (2004).

(3) S.R. Needham and T.R. Wehr, LCGC Eur. 14,

244–249 (2001).

(4) Y. Lars and H.S. Honore, J. Chromatogr., A

1020, 59–67 (2003).

(5) D.A. McLoughlin, T.V. Olah, and J.D. Gilbert,

J. Pharm. Biomed. Anal. 15, 1893–1901 (1997).

(6) S.H. Hoke II, J.A. Tomlinson, R.D. Bolden,

K.L. Morand, J.D. Pinkston, and K.R.

Wehmeyer, Anal. Chem. 73, 3083–3088 (2001).

(7) Y.H. Zhang, X.Y. Gong, H.M. Zhang, R.C.

Larock, and E.S. Yeung, J. Comb. Chem. 2,

450–452 (2000).

(8) C. Zhou, Y. Jin, J.R. Kenseth, M. Stella, K.R.

Wehmeyer and W.R. Heineman, J. Pharmac.

Sci. 94, 576-589 (2005).

(9) A. Rajan, J. mullen, N. Bhatnagar, A. Dubey, A.

Niemz, B. Chakravarti, and D.N. Chakravarti,

JALA 9, 312–317 (2004).

(10) P. Patel, S. Osechinskiy, J. Koehler, L. Zhang, S.

Vajjhala, C. Phillips, and S. Hobbs, JALA 9,

185–191 (2004).

(11) N. Barbarin, D.B. Mawhinney, R. Black, and J.

Henion, J. Chromatogr., B 783, 73–83 (2003).

(12) K. Cabrera, J. Sep. Sci. 27, 843–852 (2004).

(13) J.E. MacNair, K.C. Lewis, and J.W. Jorgenson,

Anal. Chem. 69, 983–989 (1997).

(14) J.E. MacNair, K.D. Patel aqnd J.W. Jorgenson,

Anal. Chem. 71, 700–708 (1999)

(15) A.D. Jerkovich, J.S. mellors and J.W. Jorgenson,

LCGC 21, 600–610 (2003).

(16) J.A. Lippert, B. Xin, N. Wu, and M.L. Lee, J.

Microcolumn Sep. 11, 631–643 (1997).

(17) N. Wu, J.A. Lippert and M.L. Lee, J. Chro-

matogr., A 911, 1–12 (2001). n

STANDARD CONDITION

Chloramphenicol Naproxen Mefenamic Acid

(50 mg/mL) (75 mg/mL) (50 mg/mL)

tr, min Peak Area tr, min Peak Area Res tr, min Peak Area Res

Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD)

0.682 535530 1.107 499236 12.85 1.545 616132 13.30(0.066) (0.37) (0.076) (0.44) (0.061) (0.029) (0.58) (0.022)*Res 5 Resolution for adjacent pairs: Chloramphenicol/Naproxen and Naproxen/Mefenamic Acid.

FAST CONDITION

Chloramphenicol Naproxen Mefenamic Acid

(50 mg/mL) (75 mg/mL) (50 mg/mL)

tr, min Peak Area tr, min Peak Area Res tr, min Peak Area Res

Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD) Avg (%RSD)

0.323 252980 0.441 232183 4.00 0.562 284657 4.10(0.28) (0.63) (0.28) (0.51) (0.17) (0.26) (0.66) (0.25)*Res 5 Resolution for adjacent pairs: Chloramphenicol/Naproxen and Naproxen/Mefenamic Acid.

Table IV*: Performance of UPLC versus gradient conditions (n = 5 injections)

CA MFA

Avg [Found], mg/mL Avg [Found], mg/mL

(%RSD) (%RSD)

Dilution Standard Fast Standard Fast

1 1.403 1.400 0.850 0.845(0.12) (0.35) (0.96) (1.22)

2 1.404 1.414 0.858 0.862(0.10) (0.26) (0.38) (0.38)

3 1.399 1.422 0.845 0.853(0.16) (0.14) (0.63) (1.03)

Average 1.402 1.412 0.851 0.853(%RSD) (0.16) (0.79) (0.75) (1.03)

Table V: Dose formulation analysis


Developing Columns for UPLC: DesignConsiderations and Recent Developments

Eric S. Grumbach, ThomasE. Wheat, Marianna Keleand Jeffrey R. Mazzeo

Address correspondence [email protected]

hromatographers are faced withthe challenge of developing sep-arations that completely charac-terize the constituents of their

sample. A new tool for meeting these chal-lenges became generally available in 2004.This new class of separation science, UltraPerformance Liquid Chromatography orUPLC, provides improved resolution, speed,and sensitivity. This is achieved through theuse of columns with very small particlepackings and a matching family of instru-ments developed simultaneously to providefull compatibility between chemistry andinstrumentation. In describing this new sep-aration power, it is essential to consider thekey parameters that influence peak resolu-tion and ultimately lead to successful chro-matographic methods.

Resolution between two chromatographicpeaks is determined by the distance betweentwo peaks relative to their respective peakwidths. The resolution equation provides aquantitative model for the three parametersthat control resolution: efficiency, selectivityand retentivity.

Retention (k or k9) and selectivity (a) are

chemical factors describing the interactionamong the analyte molecules, the mobilephase, and the stationary phase. In contrast,

efficiency (N or plate count) describes thephysical process of band-broadening duringthe separation. Developing a chromato-graphic method is based upon the systematicmanipulation of these three parameters.

Most method development strategiesfocus on retention and selectivity becausethey are easy and economical to manipulate.Resolution is improved by increasing theretention (k) of all of the peaks. Increasingretention, however, increases peak width,resulting in lower sensitivity, and reducessample throughput. Selectivity describes theelution sequence of the peaks relative to oneanother, that is, relative retention. It can bemanipulated by several parameters includingmobile phase pH, organic modifier, andbonded phase.

Efficiency is less often used to improve aseparation because it is difficult to changeexperimentally and because any improve-ments only contribute to resolution as thesquare root. Efficiency, however, can be sig-nificantly improved by reducing the diame-ter of the particle. A column packed with1.7 mm particles would offer a 1.7 foldimprovement in resolution compared to acolumn packed with 5 mm material. Thisresolution increase is defined by narrower,lower volume peaks so sensitivity is alsoincreased. This paper will focus on the chal-lenges of improving resolution and effi-ciency by utilizing highly efficient 1.7 mmparticle packed columns. The requirementsinclude the design and development of the

The challenges in developing a new chromatographic particle for UPLC™separations are described. Columns packed with this new UPLC particlemust meet or exceed reproducibility and longevity expected for othermodern HPLC columns under conditions that are more mechanically andchemically demanding. Additionally, several new bonded phases provideflexibility for methods development, enabling the introduction of newproducts to be brought to market faster.

C

+

-

=1

1

4 k

kNRs

αα


base chromatographic particle, the prepara-tion of reproducible columns with maxi-mized life using these particles, and the pro-vision of modern reversed-phase selectivitywith different stationary phase ligands.

New Particle TechnologyThe use of smaller particle packing materialsincreases resistance to flow so that the

columns operate at higher backpressure. Inaddition, the optimal linear velocity formaximum separation efficiency of a 1.7 mmparticle necessitates operation at higher flowrates, generating even higher pressures,sometimes as high as 15,000 psi. Silica-based materials do not possess the mechani-cal strength or efficiency necessary to meetthe demands of UPLC separations. The def-inition of a new particle to meet theserequirements must, therefore, includeimproved physical stability. This strengthmust be achieved without compromising themass loading capacity of the material that isrelated to the large surface area producedwith fully porous packing material. The newmaterial must also be stable to a wide rangeof chemical operating conditions while min-imizing any secondary or mixed mode inter-actions with a wide range of analytes.

A new bridged ethylsiloxane/silica (BEH)hybrid particle was synthesized to meet thesedemands (Figure 1). It provides improvedmechanical strength even when formed intofully porous particles. The narrow size distri-bution of the particles facilitates packinginto high efficiency columns. The organic-

Pore Diameter* 130 ÅPore Volume* 0.7 mL/gSurface Area* 185 m2/g90/10 Ratio* 1.5*Expected or approximate values.

Table I: BEH particle

ACQUITY UPLC C18 C8 Shield RP18 Phenyl

BEH Chemistry

Ligand Type Trifunctional C18 Trifunctional C8 Monofunctional TrifunctionalEmbedded Polar C6 PhenylGroup

Ligand Density 3.1 mmol/m2 3.2 mmol/m2 3.3 mmol/m2 3.0 mmol/m2

Carbon Load 18% 13% 17% 15%Endcap Style Proprietary Proprietary TMS ProprietarypH Range 1–12 1–12 2–12 2–12

Table II

30 °C 7700 psi60 °C11,500 psi

Injection

1000

500

1

1.50 3.00 4.50 1.50 3.00 4.50Time (min) Time (min)

Figure 1: Bridged ethylsiloxane/silica hybrid particle provides improved mechanical and chemical stability for UPLC separations.

Figure 2: Mechanical and chemical aging study over 1000 injections at pH 11.3. Column:ACQUITY UPLC BEH C18 2.1 3 50 mm, 1.7 mm. Conditions: acetonitrile-methyl pyrrolidine buffer,45:55 (v/v) at 0.9 mL/min. Test probes: butyrophenone, protriptyline and amitriptyline.


inorganic hybrid, with ethylsiloxane bridgesboth on the surface and throughout thebody of the material, provides a broaderrange of chemical stability, especially the pHoperating range (pH 1–12), while minimiz-ing interactions of the matrix with any ana-lyte functionalities. The properties of thispacking material are summarized in Table I.These characteristics are typical of modernreversed-phase HPLC packings. The slightlylarger pore size improves the accessibility forlarger analytes, while the surface area isabout the same as first generation methylhybrid packings. The particle size distribu-tion is among the narrowest of modernpackings.

Column Lifetime Columns packed with this new UPLC parti-cle must meet or exceed the reproducibilityand longevity expected for other modernHPLC columns. Column lifetime is a broadterm that reflects both physical and chemicalchanges to the packing as well as the adsorp-tion of sample components. Chemical sta-bility depends primarily on the effect ofmobile phase pH and solvent selection. Thehigher operating pressures associated withsub-2 mm particle packed column could also

Time (min)0.20 0.40 0.60 0.80 1.00 1.20 1.40

Injection 19000 psi

Pc = 44

Injection 22509000 psi

Pc = 43A

ten

olo

l

Met

op

rolo

l

Pin

do

lol

Ate

no

lol

Met

op

rolo

l

Pin

do

lol

Figure 3: Column stability maintained over 2200 analyses at pH 2.0 with protein precipitated rat plasma samples. Column: ACQUITY UPLC BEH C182.1 3 50 mm, 1.7 mm. Mobile phase A: 0.1% triflouroacetic acid in water, mobile phase B: 0.08% triflouroacetic acid in acetonitrile. Gradient from10–40% B over 1 minute curve 7, 40–90% B from 1.0 to 1.1 minutes, hold for 0.4 min, 95–10% B from 1.5–1.6 minutes, hold for 0.5 min. Flow rate0.7 mL/min, 5.0 mL injection; temperature 30 °C; detection UV at 272 nm.

Figure 4: Selectivity difference between C18 and C8 alkyl chain columns. Column: ACQUITY UPLCBEH C18 and C8 2.1 3 100 mm, 1.7 mm. Isocratic 28% methanol; at flow rate 0.5 mL/min; temper-ature 50 °C; 5.0 mL injection; detection UV at 254 nm. Analytes: 1 HMX, 2 RDX, 3 1,3,5-TNB, 4 1,3-DNB, 5 NB, 6 Tetryl, 7 TNT, 8 2-Am-4,6-DNT, 9 4-Am-2,6-DNT, 10 2,4-DNT, 11 2,6-DNT, 12 2-NT, 13 4-NT, 14 3-NT.

1

3

2

9

13

8

12

6

11

5

4

14

C18

C8

Time (min)

0.00 2.00 4.00 6.00 8.00 10.00

1

3

2

4

9

13

8

12

6

11

5

14


compromise bed stability. All of these phe-nomena are accelerated at elevated tempera-tures. The physical and chemical effects ofmobile phase extremes were examined asshown in Figure 2. Both the mechanical(pressure tolerance) and chemical (pH andtemperature) stability were measured asretention and efficiency, with no losses over2000 injections. The same long term stabil-ity and performance is observed when usingcomplex sample matrices shown in Figure 3.In this case, protein precipitated rat plasmawas injected after evaporation and reconsti-tution in the initial mobile phase conditions.The column maintained initial peak capac-ity and selectivity for over 2200 injections.While it is never possible to predict columnlife absolutely for all combinations of sampleand operating conditions, these experimentsare consistent with this new UPLC particlemeeting or exceeding the number of injec-tions expected for traditional HPLCcolumns.

Column SelectivityA 1.7 mm particle packed column providessignificant improvements in resolutionbecause efficiency is better. Separation of the

Figure 5: Selectivity difference between C18 and embedded polar group columns for phenolic compounds. Column: ACQUITY UPLC BEH C18 andShield RP18 2.1 3 50 mm, 1.7 mm. Isocratic 45% methanol with 0.1% formic acid; at flow rate 0.6 mL/min; temperature 40 °C; 5.0 mL injection; detec-tion UV at 270 nm. Analytes: 1 quercetin, 2 kaempferol, 3 isorhamnetin.

Figure 6: Similar selectivity is often observed between C18 and embedded polar group. Isoe-lutropic conditions are used to distinguish changes in retentivity and selectivity. Column: ACQUITYUPLC BEH C18 and Shield RP18 2.1 3 100 mm, 1.7 mm Mobile phase A: 20 mM ammonium bicar-bonate pH 10.0, mobile phase B: acetonitrile; at flow rate 0.5 mL/min; temperature 30 °C; 10.0 mLinjection; detection UV at 210 nm. Analytes: forced degradation of terbinafine HCl by 8.0 Nhydrochloric acid.

Time (min)

0.00 1.00 2.00 3.00 4.00 5.00 6.00

1 2

1

2

C18

Shield RP18

Time (min)

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00

1

Terbinafine

Terbinafine

C1865% Acetonitrile

Shield RP18

65% Acetonitrile


components of a sample, however, stillrequires a bonded phase that provides bothretention and selectivity. Four bondedphases are available for UPLC separations:ACQUITY UPLC™ BEH C18 and C8(straight chain alkyl columns), ACQUITYUPLC BEH Shield RP18 (embedded polargroup column), and ACQUITY UPLCBEH Phenyl (phenyl group tethered to thesilyl functionality with a C6 alkyl). Thecharacteristics of these stationary phases aresummarized in Table II. Each provides a dif-ferent combination of hydrophobicity,silanol activity, hydrolytic stability, andchemical interaction with the analytes. Theeffect of these properties on separations canbe described briefly.

The C18 and C8 UPLC columns havealkyl chain bonded phases trifunctionallybonded to the particle surface to ensure thebest hydrolytic stability. Compared to theC18 column, the shorter chain length C8

bonded phase is less hydrophobic, and,therefore, less retentive in general. Althoughselectivity differences seldom result fromchain length differences, changes in peakelution order can occur. As shown in Figure4, a set of 14 nitroaromatic compounds wereanalyzed on both C18 and C8 stationaryphases. Baseline resolution was achieved onboth stationary phases. However, less reten-tivity is observed on the C8 column. Thereis also a change in the elution order for thepeaks.

An embedded polar group column canexhibit significantly different selectivitycompared to linear alkanes (1,2). TheACQUITY UPLC BEH Shield RP18 col-umn includes an embedded carbamategroup that shows preferential retention ofhydrogen-bond donors. Figure 5 demon-strates the selectivity differences between thestraight chain alkyl C18 and the embeddedpolar group column for a set of flavanoids.

For analytes that do not specifically interactwith the embedded polar group, the columnbehaves as a shorter chain length alkyl col-umn, as shown in Figure 6. The embeddedpolar functionality also suppresses surfacesilanol activity, reducing peak tailing, espe-cially for basic analytes. Finally, the embed-ded polar group provides compatibility withhighly aqueous mobile phases. The embed-ded carbamate group allows the stationaryphase to resist pore dewetting by increasingthe water concentration at the surface layerof the pores. The combination of the charac-teristics of ligands with embedded polargroups provides unique features, mostimportantly, an alternative selectivity toalkyl ligands.

Columns with phenyl ligands provideanother alternate selectivity. Due to the p–pbonding orbital interactions, phenylcolumns provide unique and specific selec-tivity with aromatic compounds and otheranalytes with similar π electrons. In Figure7, the separation of nonsteroidal anti-inflammatory drugs on the phenyl column iscompared to that on a C18 column. Theselectivity differences can be magnified bychanging the organic modifier from acetoni-trile to methanol, increasing the retention ofp-acids (2).

ConclusionThe performance barriers of traditionalchromatographic packings have beenremoved with the development of a new,highly efficient, mechanically strong,1.7 mm bridged ethylsiloxane/silica hybridparticles developed specifically to meet thechallenge of routinely applying UPLC in themodern chromatographic laboratory. Theseparticles can be packed in columns that meetor exceed the lifetimes expected for modernHPLC columns. Stability over a broad pHoperating range combined with the severalavailable bonded phases provide flexibilityfor methods development. This flexibilityenables methods development to be moreefficient, allowing products to be brought tomarket faster. The power of these ultra-effi-cient columns is combined with a low dis-persion Ultra Performance LC™ system tosuccessfully transfer existing HPLC methodsor to develop new, fast chromatographicmethods that offer substantial improve-ments in resolution, sensitivity and samplethroughput.

Time (min)

0.00 1.00 2.00 3.00 4.00 5.00 6.00

C18

Phenyl

1

1

2

4

3

5

2

4

35

Figure 7: Selectivity difference between C18 and phenyl columns for aromatic compounds. Col-umn: ACQUITY UPLC BEH C18 and Phenyl 2.1 3 50 mm, 1.7 µm. Isocratic 55% methanol with 0.1%formic acid; at flow rate 0.5 mL/min; temperature 30 °C; 5.0 mL injection; detection UV at 254 nm.Analytes: 1. suprofen, 2. tolmetin, 3. naproxen, 4. fenoprofen, 5. diclofenac, 6. ibuprofen.


nungen ohne Kompromisse,” M.E. Swartz, D.Sievers, Labo 36 (3) (2005), in press

“UPLC-MS bei in-vitro-Metabolisierungsstudienvon Wirkstoffen”, M.E. Swartz, C. Thomas,LaborPraxis, 29(3) (2005), in press

“Selektivität in der Ultra-Performance Chromtogra-phie”, U.D. Neue, E.S. Grumbach, D. Sievers,Labo, 36(4) 2005, submitted for publication.

“Sprengstoffanalyse mit der UPLC: Analysen-geschwindigkeit und Säulenselektivität”, U.D.Neue, E.S. Grumbach, D. Sievers, NachrichtenChem. 51(5) (2005), submitted for publication.

White Papers, Brochures, and Product Bulletins

ACQUITY UPLC System, Brochure (2004).720000820EN.

UPLC: New Boundaries for the ChromatographyLaboratory, White Paper (2004).720000819EN.

Ultra Performance LC by Design, Brochure (2004).720000880EN.

Waters ACQUITY UPLC FlexCart, Product Bulletin(2005). 720001089EN.

Waters ACQUITY UPLC Sample Organizer, Prod-uct Bulletin (2005). 720001088EN.

Connections INSIGHT Service for the ACQUITYUPLC System, Brochure (2005).720001019EN.

Posters

“Ultra Performance Liquid Chromatography: Fasterand Higher Resolution Peptide Maps,” J.R.Mazzeo, P.D. Rainville, T.E. Wheat, E.S. Grum-bach, and D.M. Diehl, (WCBP 2005).WA41841.

“Practical Limits of High Performance Liquid Chro-matography — Current Status,” M. Kele, U.Neue, K. Wyndham, J. Mazzeo, and T. Walter(2005). WA40161.

“Increased Throughput and Sensitivity Obtainedwith 1.0 mm i.d. Columns Packed with 1.7 mmParticles for High Resolution, Ultra Fast Stabil-ity-Indicating Method Development,” E.S.Grumbach, T.E. Wheat, J.R. Mazzeo, and D.M.Diehl, (AAPS 2004). WA40534.

“High Speed Explosives Monitoring Using UPLC,”M. Benvenuti, A. Aubin, J. Romano, and J.Krol, (2004). 720000950EN.

“A Rapid Automated Approach to the Generationand Visualization of In Vitro Metabolism, Sol-ubility, and Log D Using LC/MS/MS andUPLC/MS/MS,” W. Potts, R. Plumb, and K.Yu, (Montreux 2004). 720001008EN.

J.R. Mazzeo, D.M. Diehl, D. Sievers, Labor-Praxis, 28 (10), 48–50 (2004).

“In the News: UPLC Detects More MetabolitesThan HPLC,” Trends Anal. Chem. 23(10–11),V, (2004).

“Ultra Performance LC: Von High zu Ultra“, U.D.Neue, M. Kele, J.R. Mazzeo, E.S. Grumbach, K.Wyndham, J. Morawski, D. Sievers, Nachrichtenaus der Chemie, 1217–1218 52(11),(2004).

“Emerging Technologies in the MS Arsenal”, M.Balogh, LCGC 22(11), 1082–1090, (2004).

“Degradation Analysis Using UPLC”, M.E. Swartz,Pharm. Formulation & Quality 6(5) 40–42(2004).

“UPLC: Chromatographie jenseits der klassischenGrenzen der HPLC“, M.E. Swartz, B.J. Mur-phy, D. Sievers, GIT Separation 22(1), 10–11,(2004).

“Evolution and Revolution in Chromatography,”Genetic Eng. News, 24(13) 46–47, (2004).

“Addressing Progress in Genomics Technologies,Highlights from the Recent ‘Genomes to Sys-tems’ Conference,” Genetic Eng. News 24(18), 1,9–10, 12, 28 (2004).

“LC Technologies Tackle Reproducibility,” R&DMag., 46 (8), 26–28, (2004).

“UPLC: Expanding the Limits of HPLC,” M.E.Swartz, B.J. Murphy, D. Sievers, G.I.T. Lab. J.,8(5), 43–45 (2004).

“Agilent and Waters Give HPLC a Makeover – Com-panies Reinvent an Old Standard of the Pro-teomics Toolbox,” The Scientist 18(16), 37(2004).

“Ultra Performance Liquid Chromatography:Tomorrow’s HPLC Technology Today,” M.E.Swartz, LabPlus Int. 18(3), 6–9 (2004).

“Analytical Systems Emphasize Automation – LatestOfferings Aim to Increase Productivity,” GeneticEng. News, 24(9), 44 (2004).

“Agilent and Waters Make Splash with New LiquidChromatography Product Intros at Pittcon,”ProteoMonitor 4(11), 1 (2004).

“Pittcon ’04: It’s All About Innovative Technologyand Applications,” Am. Lab. 36(8), 12, 14–15(2004).

“Pittcon 2004: An Upbeat Business Climate andInnovative Products Were Features of This Year’sShow,” Chem. Eng. News 82(13), 25–28 (2004).

“New Product Standouts at Pittcon 2004:ACQUITY UPLC,” Instr. Bus. Outlook 12(3),6,9 (2004).

“UPLC - Ultraschnelle chromatographische Tren-

Waters ACQUITY UPLC SystemLiterature and PresentationReferences

Peer-Reviewed Publications

“UPLC: An Introduction and Review”, M.E. Swartz,J. Liq. Chrom. 28(7,8), 1253–1263 (2005).

“Increasing throughput and information content forin vitro drug metabolism experiments usingultra-performance liquid chromatography cou-pled to a quadrupole time-of-flight mass spec-trometer,” J. Castro-Perez, R. Plumb, J.H.Granger, I. Beattie, K. Joncour, and A. Wright,Rapid Commun. Mass Spectrom. 19, 843–848(2005).

“HPLC-MS-based methods for the study of metabo-nomics,” I.D. Wilson, R. Plumb, J. Granger, H.Major, R. Williams, and E.M. Lenz, J. Chro-matogr., B 817(1), 67–76 (2005).

“Ultra Performance Liquid Chromatography Cou-pled to Quadrupole-Orthogonal Time-of-FlightMass Spectrometry,” R. Plumb, J. Castro-Perez,J. Granger, I. Beattie, K. Joncour, and A.Wright, Rapid Commun. Mass Spectrom. 18,2331–2337 (2004).

“High Resolution “Ultra Performance” Liquid Chro-matography Coupled to oa-TOF Mass Spec-trometry as a Tool for Differential MetabolicPathway Profiling in Functional Genomic Stud-ies,” I.D. Wilson, J.K. Nicholson, J. Castro-Perez, J.H. Granger, K.A. Johnson, B.W. Smith,and R.S. Plumb, J. Proteom. Sci., in press.

“Ultra Performance Liquid Chromatography,” J.R.Mazzeo, U.D. Neue, M. Kele, R. Plumb, Anal.Chem., submitted for publication.

Articles

“New Frontiers in Chromatography,” M. Swartz andB. Murphy, Am. Lab. 37(3), 22–27 (2005).

“UPLC-Trennung von PAKs und Sprengstoffen”,M.E. Swartz, C.Weis, GIT-Labor-Fachzeitschrift,49(3), 243–244 (2005).

“Waters’ Triumph of the Tiny: 1.7-mm particlechemistry edges company forward in the raceto supply the most powerful LC equipment”, J.Russell, Bio-IT World 3(12), 14–15(2004).

“UPLC: Expanding the Limits of HPLC, M.E.Swartz, B.J. Murphy, D. Sievers, Screening 5(3),36–38 (2004).

“Kicking Separation Up A Notch, Ultra-high-pres-sure chromatography improves speed, resolu-tion, and sensitivity of HPLC,” C.M. Henry,C&EN News 82(47), 68, 70–71 (2004).

“Ultra Performance LC: Neue Dimension in derFlussigkeitschromatographie, ” E.S. Grumbach,

BIBLIOGRAPHYBIBLIOGRAPHY


“The Application of Sub-2 mm Porous Chromato-graphic Particles with Operating Pressures up to15,000 PSI for Analysis of Pharmaceutical Com-pounds in Biological Fluids,” W. Potts, R.Plumb, J. Castro-Perez, and K. Yu, (Montreux2004). 720001010EN.

“Evaluation of UPLC/MS/MS for the Validated Bio-analytical Quantification of Amitriptyline inHuman Plasma,” T. Khan, C. Cuppett, S.Preece, and L. Southern, (AAPS 2004).720001012EN.

“Ultra Performance Liquid Chromatography (UPLC)Coupled With OA-TOF-MS for MetaboliteIdentification,” M. McCullagh, L. Southern, C.Duckett, H. Major, J. Castro Perez, J. Lindon3,J. Nicholson, I. Wilson, Waters Corporation,AstraZeneca, Macclesfield, U.K., and ImperialCollege of Science Technology and Medicine,London, U.K. (Montreaux 2004).720001004EN.

“Analysis of Complex Natural Products Using UltraPerformance Liquid Chromatography CoupledWith Time-of-Flight Mass Spectrometry,” M.McCullagh, L. Southern, C.A.M. Pereira, andJ.H. Yariwake (Montreaux 2004).720001007EN.

“Application of ESCi‘ UPLC‘ MS/MS in Drug Dis-covery and Development by ACQUITY UPLC‘with Quattro Premier,‘” K. Yu, P. Alden, S. Li,and E. Kerns (CPSA 2004 and Montreaux2004). 720000998EN.

“Column Performance Prediction and Comparison,”U. Neue and M. Gilar (2004). WA40164.

“UPLC‘ with MS (TOF): New Separations Technol-ogy for High Throughput Metabonomics,” J.Granger, R. Plumb, J. Castro-Perez, R. Williams,E.M. Lenz, and I.D. Wilson (ISSX 2004).720000992EN.

“Advances in Automated Metabolite IdentificationSoftware Tools Coupled with In-Silico Metabo-lite Prediction using LC-TOF-MS-MS,” A.Baker, J. Granger, J. Castro-Perez, J. Kirby, M.McCullagh, and N. McSweeney (ISSX 2004).720000973EN.

“A Novel Ultra-HTS Method by UPLC-MS forMetabolite Detection and Identification”, J. Cas-tro-Perez, J. Granger, A. Baker, R. Plumb, I.Beattie, K. Joncour, and A. Wright (ISSX 2004),720000974EN.

“Optimizing Throughput of Physicochemical Prop-erty Analyses,” P.M. Lefebvre, D. Shave, W.B.Potts III, and R. Plumb, (2004). 720000945EN.

“UPLC Coupled with TOF MS for Complex Multi-Component Analyses,” M. McCullagh, H.Major, J. Castro Perez, L. Southern, I. Wilson, C.Duckett, J. Lindon, and J. Nicholson (2004).720000963EN.

“Utilizing Sub-2 mm Particulate HPLC Columns for

High Resolution, Ultra Fast ChromatographicMethods,” E.S. Grumbach, D.M. Diehl, andJ.R. Mazzeo (HPLC 2004). WA40513.

“Practical Limits of High Performance Liquid Chro-matography — Current Status,” M. Kele, U.Neue, K. Wyndham, J. Mazzeo, and T. Walter(HPLC 2004). WA401161.

“Column Performance Prediction and Comparison,”U. Neue and M. Gilar (HPLC 2004).WA40164.

“NanoLC/MS in Quantitative and Qualitative Pro-teomics,” I. Kass, S. A. Cohen, J. W. Finch, H.Liu, G. Gerhardt, and K. Fadgen (2004).720001020EN.

“1.0 mm i.d. Columns Packed with 1.7 µm ParticlesImprove LC/ESI-MS Limits of Detection,” E.S.Grumbach, D.M. Diehl and J.R. Mazzeo(ASMS 2004). WA40496.

“A Novel Approach to Metabolite Identification byHigh Resolution Chromatography Coupled toHigh Resolution TOF-MS,” J. Castro-Perez, R.Plumb, J. Granger, I. Beattie, K. Joncour, andA. Wright (ASMS 2004).

“HT Quantification Analysis for A Five Drug Mix-ture in Rat Plasma – A Comparison ofHPLC/MS/MS and UPLC™/MS/MS” K. Yu,D. Little, and R. Plumb (ASMS 2004).

“Intelligent Chip: A Novel Way to Track ColumnHistory”, J. Pippitt, T. Ciolkosz, R. Gilman, D.Prentice, M. Kele, E. Woods, J. Mazzeo, J.Heden, B. Smith (Pittcon 2004). WA31828.

Application and Technical Notes

“A New Paradigm for Metabolism Studies: UPLC‘/Q-

Tof,‘” J. Castro-Perez, R. Plumb, and J. Granger,(2004). 720000953EN.

“Maximizing Chromatographic Resolution ofMetabolites Using UPLC,” J. Castro-Perez, R.Plumb, J. Granger (2004). 720001016EN.

“ACQUITY UPLC For The Rapid Analysis of SoftDrinks,” Andrew Aubin, (2005).720001053EN.

“UPLC-oaTOF MS With One-Minute SeparationTimes Applied to a Metabonomics 90-Day ToxStudy,” J.H. Granger, J.N. Haselden, M.L. Beau-mont, M. Hodson, and R.S. Plumb (2005).720001054EN.

“Metabonomics Analysis Of Zucker Rat Urine UsingUPLC/MS(TOF): A Time-Course Study,” I.D.Wilson1, R. Williams, J.H. Granger, and R.S.Plumb (2005). 720001057EN.

“Non-Mammalian Metabonomics: An Analysis OfCoffee Bean Extracts”, J.H. Granger, R.S.Plumb, and J.N. Willis (2005). 720001058EN.

“The Metabolism Of Acetaminophen: HarnessingThe Power Of UPLC/MS” J.H. Granger andR.S. Plumb, (2005). 720001059EN.

“HT Quantitative Analysis For A Drug Mixture ByLC/MS/MS: UPLC/MS/MS andHPLC/MS/MS Compared,” K. Yu, D. Little,and R. Plumb (2005). 720001120EN.

“Comparison Of One- And Two-MillimeterACQUITY UPLC Columns For LC/MS”, P.Rainville, I. D. Wilson, R.S. Plumb, and K.A.Johnson (2005). 720001126EN.

BIBLIOGRAPHYBIBLIOGRAPHYContinued

NOTE: Titles with a reference number (72000xxxxEN, or

WAxxxxx) are available as .pdf files on the Waters web site at:

http://www.waters.com/uplc.

Waters, ACQUITY Ultra Performance LC, ACQUITY UPLC, UPLC, CapLC, Alliance,eCord, Micromass, and XTerra are trademarks of the Waters Corporation. All other trade-marks are the property of their respective owners.

Better collaboration. Greater productivity. Faster time to market. When Waters® Laboratory Informatics Suite is at the core of your lab, it’s all a reality.Waters Laboratory Informatics Suite – including eLab Notebook™, NuGenesis® SDMS, Empower™ and MassLynx™ – brings together all the information management solutions you need to compile and share data, improve effectiveness and make the most knowledgeable decisions possible. And it’s all integrated into one consistent application environment. Whether you’re searching for application-specific solutions or looking to completely overhaul your data management strategy, the Waters Laboratory Informatics Suite makes productivity into an exact science. Learn more at www.waters.com/informatics

© 2005 Waters Corporation Waters, eLab Notebook, NuGenesis, Empower and MassLynx are trademarks of Waters Corporation.

Together, your possibilities are endless. Together, your possibilities are endless.

Find out about Inform 2005, May 24 - 26, 2005 • Visit www.inform2005.com

CHEMISTRY CHROMATOGRAPHY INFORMATICS MASS SPECTROMETRY SERVICES

SALES OFFICES:

AUSTRIA 43 1 877 18 07

AUSTRALIA 61 2 9933 1777

BELGIUM AND LUXEMBOURG 32 2 726 1000

BRAZIL 55 11 5543 7788

CANADA 800 252 4752 X2205

CZECH REPUBLIC 420 2 617 11384

DENMARK 45 46 59 8080

FINLAND 358 09 506 4140

FRANCE 33 1 3048 7200

GERMANY 49 6196 400600

HONG KONG 852 29 64 1800

HUNGARY 36 1 350 5086

INDIA 91 80 2837 1900

IRELAND 353 1 448 1500

ITALY 39 02 27 4211

JAPAN 81 3 3471 7191

KOREA 82 2 820 2700

MEXICO 52 55 5524 7636

THE NETHERLANDS 31 76 508 7200

NORWAY 47 6 384 6050

PEOPLE’S REPUBLIC OF CHINA 86 10 8451 8918

POLAND 48 22 833 4400

PUERTO RICO 787 747 8445

RUSSIA/CIS 7 095 931 9193

SINGAPORE 65 6278 7997

SPAIN 34 93 600 9300

SWEDEN 46 8 555 11 500

SWITZERLAND 41 62 889 2030

TAIWAN 886 2 2543 1898

UK 44 208 238 6100

USA 508 478 2000 • 800 252 4752

©2005 Waters Corporation. Waters is a trademark of Waters Corporation.

720001136EN

separation science redefined - waters · ultra performance liquid chromatography coupled to...

Documents