APCI Interface for LC- and SEC-MS Analysis ofSynthetic Polymers: Advantages and Limits

Bernard Desmazieres, William Buchmann,* Peran Terrier, and Jeanine Tortajada

Laboratoire Analyse et Modelisation pour la Biologie et l’Environnement, Universite d’Evry-Val d’Essonne,CNRS UMR 8587, Bat. Maupertuis, Bd. F. Mitterrand, 91025 Evry Cedex, France

The main advantage of the APCI interface for the LC-MSanalysis of synthetic polymers resides in its compatibilitywith the main chromatographic modes: reversed-phaseliquid chromatography, normal-phase liquid chromatog-raphy, and size exclusion chromatography in organicphase, with the usual flow rates. Moreover, APCI can beused in positive or negative modes. Representative ap-plications are described to highlight benefits and limita-tions of the LC-APCI-MS technique with the analysis ofindustrial polymers up to molecular masses of 5 kDa:polyethers; polysiloxanes; and copolymers of siloxanes.Results are discussed in regard to those obtained by moreclassical techniques: SEC and MALDI-MS. The use of anAPCI interface in LC-MS and SEC-MS coupling appliedto synthetic polymers is efficient up to 2000-4500 Da.The main drawback of the APCI interface is the in-sourcedecomposition that is observed above m/z ) 2000-3000and can induce an underestimation of average molecularweights. However, APCI allows detection on a wide rangeof polarity of sample/solvent and appears to be comple-mentary to ESI.

Mass spectrometry (MS), with the introduction and thedevelopment of soft ionization methods such as electrosprayionization (ESI), by Fenn et al.,1 and matrix-assisted laser desorp-tion/ionization (MALDI), by Tanaka et al.2 and by Karas andHillenkamp,3 has been shown to be a very powerful tool forpolymer analysis.4-6 MALDI-MS allows the checking of the natureof the repeat units and the end groups, estimating the averagemolecular weights, the polymerization degrees, and the polydis-persity indices when polymers are few disperse. ESI-MS has alsobeen shown to be a very useful ionization technique also, but ESI-MS has been seldom used comparatively to MALDI-MS forpolymer analysis due to the presence of several charge states.Charge-state distribution overlaps with oligomer chain lengthdistribution; thus, very complex mass spectra can be obtained with

higher masses. Amazingly, the use of atmospheric pressurechemical ionization (APCI) as an ionization method for syntheticpolymer analysis has been scarce whereas singly charged pseudo-molecular ions are generally obtained in contrast to ESI. APCIwas initially introduced by Horning et al.7 in the mid-1970s as aninterface between high-performance liquid chromatography (HPLC)and MS. In APCI, ions are generated at atmospheric pressure,from the LC effluent through a heated pneumatic nebulizer viacorona discharge ionization. Usually, APCI is preferred over ESIfor the analysis of compounds of low polarity and of low molecularweights. During the APCI process, the mobile phase acts as areactant gas to ensure ionization. APCI is known to be somewhatless mild than ESI due to the fact that the evaporation of mobilephase is supported by a heated nebulizer (350-400 °C). To ourknowledge, only a few papers deal with the use of APCI for LC-MS polymer analysis. Huang and Rood8 showed the advantagesof APCI-MS in the infusion mode over GC/CH4C-MS with varioussilylated and unsilylated ethoxylates and carboxylates. Jandera andco-workers9,10 studied the chromatographic behavior of ethoxy-lated alcohols and ethylene oxide/propylene oxide copolymersin normal- and reversed-phase HPLC-MS using APCI-MS as adetector. Cyclic oligomers of poly(ethylene terephtalate)11-13 andpoly(butylene terephtalate)14 were characterized by reversed-phaseHPLC-APCI MS. In these reports, the use of APCI was alwaysdescribed in the positive ion mode only and over a limited massrange (usually <1500 Da). However, packed-column supercriticalfluid chromatography (SFC) with APCI detection was successfullyused in the negative ion mode to analyze phenolic oligomers byDavidson et al.15 Due to the acidity of the phenolic OH group,negative-ion APCI was ideal for providing molecular weightinformation without extensive fragmentation. Capillary SFC usingAPCI-MS as detector was used to characterize siloxane-

* To whom correspondence should addressed. E-mail: william.buchmann@univ-evry.fr.(1) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science

1989, 246, 64-71.(2) Tanaka, K. Angew. Chem., Int.l Ed. 2003, 42, 3860-3870.(3) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301.(4) Pasch, H.; Shrepp, W. In MALDI-TOF Mass Spectrometry of Synthetic

Polymers; Barth, H. G., Pash, H., Eds.; Springer-Verlag: Berlin, 2003(5) Montaudo, M. S. Mass Spectrom. Rev. 2002, 21, 108-144.(6) Montaudo, G.; Lattimer, R. P. In Mass Spectrometry of Polymers; Montaudo,

G., Lattimer, R, P., Eds.; CRC Press: Boca Raton, FA, 2002.

(7) Carroll, D. I.; Dzidic, I.; Stillwell, R. N.; Haegele, K. D.; Horning, E. C. Anal.Chem. 1975, 47, 2369-2373.

(8) Huang, S. K.; Rood, M. H. Rapid Commun. Mass Spectrom. 1999, 13, 1152-1158.

(9) Jandera, P.; Holcapek, M.; Theodoridis, G. J. Chromatogr., A 1998, 813,299-311.

(10) Jandera, P.; Holcapek, M.; Kolarova, L. J. Chromatogr., A 2000, 869, 65-84.

(11) Barnes, K. A.; Damant, A. P.; Startin, J. R.; Castle, L. J. Chromatogr., A 1995,712, 191-199.

(12) Harrison, A. G.; Taylor, M. J.; Scrivens, J. H.; Yates, H. Polymer 1997, 38,2549-2555.

(13) Bryant, J. J. L.; Semlyen, J. A. Polymer 1997, 38, 2475-2482.(14) Bryant, J. J. L.; Semlyen, J. A. Polymer 1997, 38, 4531-4537.(15) Carrott, M. J.; Davidson, G. Analyst 1999, 124, 993-997.

Anal. Chem. 2008, 80, 783-792

10.1021/ac0715367 CCC: $40.75 © 2008 American Chemical Society Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 783Published on Web 12/28/2007

polyethylene copolymers.16 Finally, compared to MALDI-MS orESI-MS, APCI technique has almost been forgotten by massspectrometrists for polymer analysis. The reason is probably thatthis ionization method is reputedly efficient for low molecularweight compounds only, since thermal decompositions are acommon feature of APCI for higher masses. Nevertheless, theactual mass limits for polymer characterization have not clearlybeen addressed due to the very small number of publications inthis field. Recently, new ionization methods have been applied topolymer analysis: atmospheric pressure MALDI17-20 (AP-MALDI),desorption/ionization on silicon21,22 (DIOS), and desorption-electrospray ionization23,24 (DESI). The main advantage of the AP-MALDI source for polymer analysis resides in its possible couplingwith analyzers enabling MSn experiments.20 With DIOS and DESItechniques, the sample preparation step is strongly simplified (ifnot suppressed). Unfortunately, these new ionization methodscannot be directly coupled with liquid chromatography, and theirapplicability remained limited to low molecular weight polymersas well. In this paper, various LC-APCI MS couplings areevaluated; the results are completed by original SEC-APCI MSexperiments. It will be shown from the results that APCI is a veryinteresting ionization technique, a serious alternative to the ESImethod for the LC-MS analysis of synthetic polymers. One of

its advantages resides in its compatibility with the main chro-matographic modes: reversed-phase liquid chromatography(RPLC), normal-phase liquid chromatography (NPLC), and sizeexclusion chromatography (SEC), with the usual flow rates. APCIoperates at flow rates of standard chromatography (typically 1 mL/min with 4.6-mm-i.d columns). Moreover, APCI can be used inpositive or negative mode. In this work, the performance of theAPCI interface will be investigated in SEC and LC-MS couplingapplied to the analysis of industrial polymers up to molecularmasses of 5 kDa: polyethers; polysiloxanes, and copolymers ofsiloxanes. Representative applications will be described to high-light benefits and limitations of the APCI technique.

EXPERIMENTAL SECTIONMaterials. Dodecanol ethoxylate C12H25-(OCH2CH2)nOH

(Mn ≈ 700 g‚mol-1), hence described as C12EOn, was purchasedfrom Sigma-Aldrich (St. Quentin-Fallavier, France). The mixturesC16/C18EOn and Empilan KM11 and KM25 polymers (Mn ≈ 700and 1500 g‚mol-1, respectively) were from Marchon (Saint Mihiel,France), and the mixture C16/C18EOn (Mn ≈ 2500 g‚mol-1) wasfrom Witco (Greenwich, CT). Hydroxyl-terminated homopolymers(poly(butylene oxide) Mn ≈ 2000 g‚mol-1 (PTHF), a mixture ofpoly(dimethylsiloxane) 5, 10, 20, and 50 cSt, Mn≈ 2500 g‚mol-1

(PDMS) and linear triblock copolymers, Mn ≈ 1900 g‚mol-1

(PDMS-b-PMS-b-PDMS) containing 33% w/w of poly(methylsilox-ane) (PMS) were obtained from Sigma-Aldrich (Saint Quentin-Fallavier, France). Solutions of polymer (with concentrationsbetween 10-4 and 10-3 M according to its nature) in the mobilephase were prepared prior to analysis. No salt was added.

LC and LC-MS Analyses. LC-MS experiments (RPLC,NPLC, SEC) were carried out using a Merck-Hitachi L6200chromatograph (Merck, Darmstadt, Germany).

RPLC analyses were carried out using C18 bonded silicacolumns (Merck Lichrospher 250 × 4 mm, dp 5 µm, WatersSymmetry, 150 × 3.9 mm, dp 4 µm). Mobile phases were chosendepending on the sample: water/acetonitrile or water/methanol(polyethers); acetonitrile/acetone (polysiloxanes). Flow rates:

(16) Just, U.; Jones, D. J.; Auerbach, R. H.; Davidson, G.; Kappler, K. J. Biochem.Biophys. Methods 2000, 43, 209-221.

(17) Doroshenko, V. M.; Laiko, V. V.; Taranenko, N. I.; Berkout, V. D.; Lee, H.S. Int. J. Mass Spectrom. 2002, 221, 39-58.

(18) Laiko, V. V.; Baldwin, M. A.; Burlingame, A. L. Anal. Chem. 2000, 72, 652-657.

(19) Creaser, C. S.; Reynolds, J. C.; Hoteling, A. J.; Nichols, W. F.; Owens, K. G.Eur. J. Mass Spectrom. 2003, 9, 33-44.

(20) Hanton, S. D.; Parees, D. M.; Zweigenbaum, J. J. Am. Soc. Mass Spectrom.2006, 17, 453-458.

(21) Lewis, W. G.; Shen, Z. X.; Finn, M. G.; Siuzdak, G. Int. J. Mass Spectrom.2003, 226, 107-116.

(22) Shen, Z. X.; Thomas, J. J.; Siuzdak, G.; Blackledge, R. D. J. Forensic Sci.2004, 49, 1028-1035.

(23) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306,471-473.

(24) Jackson, A. T.; Williams, J. P.; Scrivens, J. H. Rapid Commun. Mass Spectrom.2006, 20, 2717-2727.

Figure 1. APCI mass spectrum (positive mode) of a mixture of surfactants.

784 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

0.7-1 mL/min. Injection volumes: 5-20 µL. Isocratic mode,gradient elution or temperature programming were selectedaccording to the samples analyzed.

NPLC analyses involved several stationary phases: diol bondedsilica (Merck Lichrospher, 250 × 4.6 mm, dp 5 µm); nitrophenylbonded silica (Macherey Nagel Nucleosil, 250 × 3 mm, dp 5 µm),Mobile phases: hexane/dichloromethane/methanol (polyethoxy-lated surfactants). Flow rate: 0.5-0.8 mL/min. Injection vol-umes: 5-20 µL. gradient elution.

SEC analyses were carried out using one or two columns, 300× 8 mm (PL gel, PS/PDVB copolymer); dp 5 µm; pore diameter1000 Å. Mobile phases were dichloromethane or THF. Theinjection volume was 20 µL (concentration range 100-500 ppm).

For all chromatographic methods, the detection during thedevelopment of LC methods (before LC-MS coupling) wascarried out by UV and differential refractometry. The connectionof liquid chromatography systems to the mass spectrometer wasensured by an APCI interface. MS experiments were carried out

using a quadrupole time-of-flight mass spectrometer (AppliedBiosystems Q-Star Pulsar, Foster City, CA) equipped with an APCIIonization source (Applied Biosystems). The heater temperaturewas 300-400 °C, needle current 2-3 µA. Focusing potential was265 V, declustering potentials 1 and 2 were respectively (1) 50-85 and (2) 15-20 V, gas pressures (1) 80 and (2) 15 psi, andcurtain gas 45 psi. N2 was used as collision gas with a 3 psipressure. Data acquisition was from m/z 200 to 500 up to 2500 to5000 according to samples in the TOF-MS mode only.

MALDI-TOF MS. Experiments were performed using a Per-septive Biosystems Voyager-DE Pro STR MALDI-TOF massspectrometer (Applied Biosystems/MDS Sciex, Foster City, CA).This instrument was equipped with a nitrogen laser (λ ) 337 nm).The mass spectrometer was operated in the positive ion reflectronmode with an accelerating potential of +20 kV. Mass spectra wererecorded with the laser intensity set just above the ionizationthreshold (2500-3000 in arbitrary units, on our instrument) toavoid fragmentation and to maximize the resolution (pulse width

Figure 2. RPLC/APCI MS (positive mode) separation of a mixture of surfactants.

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 785

3 ns). Typically, 10 acquisitions corresponding to 50 shots eachwere summed for each deposit to obtain representative massspectra. External calibration using a mixture of two poly(ethyleneglycol) standards (Mh n )1000 and Mh n ) 2000) was performed withthe same 2,5-dihydroxybenzoic acid (DHB) matrix as in theexperiment. DHB matrix was purchased from Aldrich Chemicaland used without further purification. For all experiments, equalvolumes of polymer solution (10-3 M in THF) and of matrixsolution (10-1 M in THF) were mixed. About 1 µL of the resultingmixture was spotted onto the sample plate and allowed to air-dryat room temperature just before MALDI analysis.

RESULTS AND DISCUSSIONAnalysis of a Mixture of Polyoxyethylenic Surfactants

(C12/C16/C18EOn, Sample I) by RPLC/APCI MS Coupling.The first example of polymer analysis by LC-MS using an APCIinterface concerns a mixture of three poly(ethylene oxides)bearing alkyl chains of different lengths C12H25O(CH2CH2O)nH,

C16H33O(CH2CH2O)nH, and C18H37O(CH2CH2O)nH. The samplewas in fact a 1/1 mixture of polydodecanol (C12EO9) and of KM11(described as a C16/C18EO11 mixture by the supplier with about2/3 of C18 chains). These compounds belong to linear alcoholethoxylates (R-(OCH2CH2)mOH), which are widely used as indus-trial nonionic surfactants.25 Commercial formulations are complexmixtures of oligomers depending on the distribution of bothethylene oxides (EO) and alkyl end groups (R). Since thephysicochemical properties of these polymers largely depend onthe ratio of the length of the hydrophobic (R) to the hydrophilic(EO) segments, efficient characterization techniques are required.Various chromatographic methods have been described to analyzethis class of compounds and other polyether derivatives byRissler,26 Trathnigg,27 and Di Corcia.28 Different techniques (two-

(25) Thetford, D. Applications of oligomeric surfactants in polymer systems. InSurfactants in Polymers, Coatings, Inks and Adhesives; Karsa, D. R., Ed.; CRCPress: Boca Raton, FL, 2003.

(26) Rissler, K. J. Chromatogr., A 1996, 742, 1-54.

Figure 3. NPLC/APCI MS (positive mode) separation of a C16EO25 surfactant.

786 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

dimensional LC, MALDI-TOF, ESI-MS, and LC-ESI-MS) have beenrecently compared.29 Among these techniques, LC-ESI-MS cou-pling allowed a faster analysis than two-dimensional LC (4 minversus 2 h). Moreover, compared to MALDI-TOF MS and ESI-MS alone, LC-ESI-MS experiments prevent overlaps of isobaricions and permit byproduct identification more easily. Figure 1shows the APCI-TOF mass spectrum of a mixture of C12/C16/C18

EOn recorded in the positive ion mode. For the sake of compari-son, Figure 2 displays the separation of the same mixture byRPLC/APCI-TOF MS (positive ion mode). In Figure 1, the overlapof some isotopic patterns makes uneasy the assignments of ionsand the accurate determination of their relative abundances. Forexample, in the expanded region of Figure 1, m/z 715 and 716can respectively correspond to (C12 EO12 + H)+ and (C14 EO11 +NH4)+, but a part of m/z 716 can also correspond to the isotopiccontribution M + 1 from (C12 EO12 + H)+. The ions at m/z 727and 728 can arise from (C16 EO11 + H)+ and (C18 EO10 + NH4)+,m/z 699 and 700 from (C14 EO11 + H)+ and (C16 EO10 + NH4)+,but overlaps can also occur. The advantage of a chromatographicseparation before MS analysis is clear: spectra are simplified; theseparation prevents possible overlaps of isotopic patterns (or of

isobaric ions) and can reduce spectral suppression effects. Forthe most complex mixtures, the separation step becomes ex-tremely useful. For the RPLC/APCI-TOF MS coupling experimentshown in Figure 2, a C18 bonded silica column was used withmethanol as the mobile phase. Under these conditions, theseparation follows the length of fatty chains; the selectivity forthe polyoxyethylenic part is suppressed. The APCI mass spectrarecorded on-line corresponding to four fractions (A, B, C, D)exhibit two main series for each fraction: (M + H)+ and (M +NH4)+ ions as shown in Figure 2. The four fractions correspondrespectively to C12/C14/C16/C18 EOn, where C14EOn is an impurity.Average molecular weights and polymerization degrees werededuced from the sum of the abundances (M + H)+ + (M +NH4)+. The four polyethers C12/C14/C16/C18 EOn present a similarDPn ≈13 (Mh n ≈ 750-850 g‚mol-1). If it is assumed that the fourpolymers are ionized with the same efficiency independently ofthe alkyl chain length, then their relative proportions are respec-tively 43, 12, 21, and 24%. These results appear in good agreementwith the expected values: 50, 0, 17, and 33%.

Analysis of the Surfactant C16EOn (Sample II) by NPLC/APCI MS Coupling. Figure 3 is an example of the specificinterest of APCI versus ESI. The ESI process needs the use of apolar solvent or polar solvent mixture (typically water/methanolor water/acetonitrile) for an efficient ionization process; thus, ESI

(27) Trathnigg, B. J. Chromatogr., A 2001, 915, 155-166.(28) Di Corcia, A. J. Chromatogr., A 1998, 794, 165-185.(29) Buchmann, W.; Nguyen, H. A.; Cheradame, H.; Morizur, J. P.; Desmazieres,

B. Chromatographia 2002, 55, 483-489.

Figure 4. SEC/APCI MS (negative mode) separation of a mixture of surfactants C18/C16EO25.

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 787

is not compatible with NPLC. On the contrary, the APCI sourcecan work whatever the polarity of the solvent. Thus, APCI can beused in NPLC with nonpolar mobile phases. In Figure 3, a sampleof C16EO25 was analyzed by NPLC-APCI MS in the positive ionmode using a diol bonded silica column with a mixture of hexane/dichloromethane/methanol as mobile phase with gradient elution(from 92/4/4 to 76/12/12 in 40 min). The main ions consist of(M + H)+ ions. The resolution was somewhat low but sufficientto identify the components and determine the average values:DPn ) 23.1, Mh n ) 1258 g‚mol-1. The average molecular weightand polymerization degree were calculated from the (M + H)+

ions by making a sum of the mass spectra over the entire elution.Analysis of the Surfactant C18EOn (Samples III) by SEC/

APCI MS Coupling. Size exclusion chromatography is certainlythe most widely used chromatographic method by polymerspecialists in order to determine the average molecular weightsof polymers. This technique is rarely used in combination withMS. Figure 4 depicts SEC separation with APCI-TOF MS detectionof a C18EO25 surfactant performed using a PL Gel (1000 Å) columnwith dichloromethane as mobile phase. In the negative mode, (M+ Cl)- ions were typically detected up to m/z ) 2200. At thebeginning of the elution (fraction A), another distribution appearedin the 700-1200 mass range. This unexpected series of ionsdisappeared progressively during the experiment when lighteroligomers were eluted. These ions were assigned to thermaldecompositions in the source since the phenomenon was growingwith the temperature increase of the APCI heater. Anotherdistribution appeared in the fractions B and C corresponding toa C16EOn byproduct. From the APCI-MS data and assumingidentical ionization efficiencies, the polymer mixture would becomposed of 85% C18 EOn (DPn ) 22.6, Mh n ) 1264 g‚mol-1) and15% C16 EOn (DPn ) 15.4, Mh n ) 921 g‚mol-1).

Analysis of the Surfactant Witco Cetalox C16/C18EOn

(Sample IV) by SEC/APCI MS Coupling. Under the sameconditions, a similar sample but with a higher average molecularweight (C18/C16EO50) was analyzed in order to assess the upperlimit of mass detection with the APCI interface. The results aregiven in Figure 5. A total of 55% of the (M + Cl)- ions correspondto C18 EOn (DPn ) 23.4, Mh n ) 1301 g‚mol-1) and 45% corre-sponds to C16 EOn (DPn ) 20.9, Mh n ) 1163 g‚mol-1). (M + Cl)-

ions were detected up to m/z ) 2500 (see fraction A), butdecomposition remained important for higher masses. Under ourconditions, 2500 atomic mass units (amu) seemed to be the upperdetection limit. The extent of decomposition can depend on manyfactors such as the following: nature of mobile phase, sourcetemperature, source pressure, and needle current. Since thequality of the separation is related to the composition of the mobilephase, the latter was not modified. However, in order to preventthermal decomposition of the oligomers in the source, the sourcetemperature was decreased from 400 to 300 °C. Unfortunately,this did not efficiently limit the decomposition but rather reducedionization efficiency. The source heater temperature was set toan optimal value of 350 °C. The effect of source pressure wasexplored also. This pressure increases with the chromatographicflow rate but can be controlled by opening the exhaust. Unfortu-nately, the main consequence obtained by modifying this param-eter was a deterioration of the chromatographic resolution whenmoving away from the optimal value of the exhaust. Last, the

thermal decomposition can depend on the nature of repeat unitof the oligomers. As it will be shown, some oligomers can be morefragile than others.

Analysis of the Poly(tetrahydrofuran) (Sample V) byRPLC/APCI MS Coupling. Poly(tetrahydrofuran) is a linear,saturated, diol polyether derived from the polymerization oftetrahydrofuran. While soluble in almost all conventional organicsolvents, it is barely soluble in water. Poly(tetrahydrofuran) istypically used as soft segment for thermoplastic elastomers andcross-linked elastomers. The LC analysis of poly(tetrahydrofuran)was optimized by Rissler et al.30 This method was adapted, THFwas replaced by acetone in the mobile phase. Figure 6 depictsthe RPLC separation of a poly(tetrahydrofuran) APCI-MS detec-tion. The separation was achieved using two C18 columns withelution gradient, from 18% acetone in acetonitrile to 100% acetone.Mass spectra of various selected fractions are also given in Figure6. Two kinds of ionic species are detected: (M + H)+ and (M +NH4)+ adducts. From A to D, isotopic patterns are more and morespread due to the isotopic contribution of carbon. However,monoisotopic ions remain easily assignable. Based on the peakintensity of the total ion current, the estimation of the averagemolecular weight and of the polymerization degree gave DPn )27.5 and Mh n ) 1995 g‚mol-1. From the total ion current and thevarious assignments A-D, it is clear that mass detection limit ishigher than in the previous experiment. The upper mass limit isabove 3000 amu. Clearly, this limit depends on the nature of theoligomers and of the mobile-phase composition.

(30) Rissler, K. Chromatographia 2004, 59, 669-675.

Figure 5. SEC/APCI MS (negative mode) separation of a mixtureof surfactants C18/C16EO50.

788 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

Analysis of a Poly(dimethylsiloxane) (Sample VI) byRPLC/APCI MS Coupling. A further example is given in Figure7 with the RPLC separation of a mixture of poly(dimethylsiloxane).This polymer is composed of straight chains terminated with twotrimethylsilyl groups. Siloxane polymers (silicones) are known tooccur in various forms (fluids, gels, etc.) for a wide variety ofapplications.31 An efficient separation was achieved using twoMerck C18 columns and applying an elution gradient, from 80%acetone in acetonitrile to 100% acetone. Mass spectra recordedduring elution show the presence of (M + NH4)+ adducts mainly.Isotopic patterns are strongly affected by the presence of silicium.Intact oligomer chains are detected here above 3500 amu. DPn

and Mh n values were estimated from the (M + NH4)+ ions bymaking a sum of the mass spectra over the entire elution and bytaking into account the isotopic patterns (DPn ) 30.3, Mh n )2403).

Analysis of a Block Copolymer of Methyl- and Dimethyl-siloxanes (Sample VII) by RPLC/APCI MS. The last example

concerns a block copolymer of methyl- and dimethylsiloxanes.Figure 8 shows the RPLC/APCI(+) separation obtained using twoC18 bonded silica columns. The structure of such a copolymer isdescribed at the top of Figure 8. The separation was achieved inthe isocratic mode with acetonitrile/acetone 66/34 as mobile-phase composition. However, columns were set in an oven witha programmed temperature (from 25 to 80 °C, 1 °C /min). (M +NH4)+ ions were detected as main series. Each chromatographicpeak consists of overlapping of combinations of methyl anddimethyl polysiloxanes with a constant number of methyl groupswithin the chain. For example, with a retention time of 20 min,oligomers contain a total of 25 methyl groups ((x + y) × 2 + z )25). When considering the x, y, z contributions of methyl anddimethyl siloxanes, it can correspond to various (x + z)/y ratios:12/1, 11/3, 10/5, or 9/7. A contour plot (Figure 9) deduced fromthe RPLC/APCI-TOF separation was produced by means of amethod described in detail elsewhere.32 This 2D plot reports theion intensities of the various oligomers as a function of the numberof methylsiloxane and dimethylsiloxane units. Only ions with m/z

(31) Kendrick, T. C.; Parbhoo, B.; White, J. W. In Siloxane Polymers andCopolymers, The Chemistry of Organic Silicon Compounds; Patai, S., Rap-poport, Z., Eds.; John Wiley: Chichester, 1989.

(32) Terrier, P.; Buchmann, W.; Cheguillaume, G.; Desmazieres, B.; Tortajada,J. Anal. Chem. 2005, 77, 3292-3300.

Figure 6. RPLC/APCI MS (positive mode) separation of a poly(tetrahydrofuran).

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 789

> 600 Da were taken into account in order to avoid thecontribution of low-mass ions coming from in-source decomposi-tions. The estimation of the average polymerization degrees gavefor each sort of repeat unit: DPn (DMS) ) 14.9, DPn (MS) ) 5.4.

Comparison of LC-APCI MS Results with SEC and MALDIMS Data. Finally, average molecular weights deduced from theLC-APCI MS experiments were compared with those obtainedby SEC and MALDI MS (see Table 1). First, MALDI MS and

Figure 7. RPLC/APCI MS (positive mode) separation of a poly(dimethylsiloxane).

Table 1. Average Molecular Weights (Mn), Average Degrees of Polymerization (DPn), and Polydispersity Indexes (Ip)Deduced from SEC, MALDI-MS, and LC-APCI MS

SEC dataa MALDI-TOF MS+datab LC-APCI-MS data

sample Mn DPn Ip Mn DPn Ip mode Mn DPn Ip

I C12EOn 529 7.8 1.17 716 12 1.08 RPLC + 745 12.7 1.02C16EOn 917c 15c 1.09c 1073 18.9 1.06 805 12.8 1.03C18EOn 1095 18.7 1.06 825 12.6 1.03

II C16EOn 1545 29.6 1.06 1596 30.8 1.02 NPLC + 1258 23.1 1.04III C18EOn 1726 33.1 1.07 1473 27.3 1.03 SEC - 1264 22.6 1.07IV C16EOn 2132c 42.6c 1.07c 2225 45 1.03 SEC - 1163 20.9 1.07

C18EOn 2253 45 1.03 1301 23.4 1.11V PTHF 2506 34.5 1.98 909 12.4 1.29 RPLC + 1995 27.5 1.3VI PDMS 2430 32.8 1.48 1127 13.04 1.04 RPLC + 2403 30.3 1.1VII PDMS-b-PMS-b-PDMS 1437 N/A 1.77 (DMS) 873 11.8 1.17 RPLC + 1107 14.9 1.08

(MS) 298 5 1.53 322 5.4 1.17

a Refractometric detection. Poly(ethylene oxides) were used as standards. b 2,5-Dihydroxybenzoic acid was used as MALDI matrix. c Thesedata are an average measurement of C16EOn + C18EOn.

790 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

LC-APCI MS provide much more information than SEC. Forexample, SEC gives only one average molecular weight value forthe mixture C16/C18EOn (samples I, IV) whereas MALDI MS andLC-APCI MS can give one value of Mn and DPn for each polymeras a function of the fatty chain length. In the case of sample VII,SEC cannot determine the composition of the copolymer. Second,the results concerning the molecular weights depend on thepolymer. With poly(ethylene oxides) (samples I-IV), lowermolecular weight values were obtained from LC-APCI MS thanfrom SEC or MALDI MS analyses (except for C12EOn). Averagemolecular weights deduced from LC-APCI-MS for sample IV seemstrongly underestimated. Fragmentations of the heavier chainscan explain the strong differences in the estimation of polymer-ization degrees. For the poly(tetrahydrofuran) (V), the PDMS (VI),

the values deduced from LC-APCI MS were higher than thosefrom MALDI-MS but slightly lower than SEC values. This can bepartially explained by a lack of desorption-ionization efficiency inMALDI-MS with the heaviest oligomers. The copolymer PDMS-b-PMS-b-PDMS (VII) leads to similar results by both MALDI-MSand LC-APCI-MS. As already mentioned, SEC does not seem tobe the most appropriate method in this case because thistechnique is not able to distinguish the different repeat units ofthe copolymer. Concerning the polydispersity indexes, except forsample I, they roughly decrease in the following order: SEC >LC-APCI MS > MALDI-MS probably because the detection ofthe heaviest chains is limited in both MALDI-MS and APCI-MSanalyses. It must be emphasized that even if LC-APCI MS canlead to an underestimation of the average molecular weights; theaccuracy of the quantitative results depends on the polymer.Owing to the separation step before mass analysis, LC-APCI MScoupling provides the best qualitative data.

To conclude, the use of APCI interface in LC-MS and SEC-MS coupling applied to synthetic polymers is efficient up to 2000-4500 Da, pseudomolecular ions are easily detected both in normal-and in reversed-phase LC conditions, as well as in organic-phaseSEC (whatever the ion mode: positive or negative). The mainresults are that the analysis of polyethers is possible in RPLC,NPLC, or SEC APCI-MS coupling up to m/z ) 2500 at least. Theanalysis of polysiloxanes is easy up to at least m/z ) 4500 in RPLCor SEC/APCI MS, in the positive or negative ion modes. The maindrawback of the APCI interface is the in-source decomposition,

Figure 8. RPLC/APCI MS (positive mode) separation of a methylsiloxane/dimethylsiloxane triblock copolymer.

Figure 9. Contour plot deduced from a RPLC/APCI MS (positivemode) separation of a methylsiloxane/dimethylsiloxane triblock co-polymer (ion intensities reported as a function of the number ofmethylsiloxane and dimethylsiloxane units).

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 791

which is observed above m/z ) 2000-3000 and can result inunderestimation of molecular weights. However, APCI allowsdetection on a wide range of polarity of sample/solvent andappears to be complementary to ESI. Fundamental studies of thechemical ionization process and a comparison with the newlydeveloped atmospheric pressure photoionization33,34 could permit

the extension of the range of detectable molecular weights byoptimizing source parameters.

Received for review July 20, 2007. Accepted October 29,2007.

AC0715367

(33) Raffaelli, A.; Saba, A. Mass Spectrom. Rev. 2003, 22, 318-331. (34) Robb, D. B.; Covey, T. R.; Bruins, A. P. Anal. Chem. 2000, 72, 3653-3659.

792 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008