D‑α-tocopheryl Polyethylene Glycol 1000 Succinate: A View fromFTICR MS and Tandem MSJuan Wei,† Anthony Bristow,‡ Eileen McBride,‡ David Kilgour,† and Peter B. O’Connor*,†

†Department of Chemistry, University of Warwick, Coventry CV4 7AL, United Kingdom‡AstraZeneca, Macclesfield SK10 4TF, United Kingdom

*S Supporting Information

ABSTRACT: D-α-tocopheryl polyethylene glycol 1000 succinate(TPGS) is an important polymeric excipient frequently used in drugformulation. However, differing compositions of the TPGS samplesbetween batches are believed to result in variable performance of theformulated product. Herein, a high performance method usingFourier-transform ion cyclotron resonance (FTICR) mass spectrom-etry (MS) and tandem mass spectrometry (MS/MS) to analyze thecomposition of TPGS samples and the structure of TPGS wasestablished. Aided by high mass accuracy and high resolution, the fullMS overview of TPGS is able to provide composition information,and diagnostic fragments from collisionally activated dissociation(CAD) and electron capture dissociation (ECD) MS/MS can be usedfor the identification of the TPGS structure. ECD and CAD showdifferent preferences in bond cleavage, and an interesting cross-ringcleavage was generated by CAD. Fragmentation information from ECD/ECD MS3 is useful for providing confidence in theresults. The influence of different ionization agents (Na+, Li+, and Ag+) on fragmentation of TPGS was investigated with thesilver adduct providing different fragments. In addition to the methodology study, the MS and MS/MS results from four batchesof TPGS samples from two manufacturers were compared. This method can be utilized for the composition and structure studyof many other polymeric compounds. FTICR MS/MS demonstrated its promising role as a structural characterization toolcomplementary to traditional spectroscopy techniques.

The substance D-α-tocopheryl polyethylene glycol 1000succinate (TPGS) (chemical structure shown in Figure

1a) is an amphiphilic polymer frequently used in drug

formulations, and was approved as a water-soluble alternativeof vitamin E.1 TPGS contains a lipophilic head, the tocopherylmoiety, also known as vitamin E, and a polar polyethyleneglycol (PEG, Mn 1000) tail, which are connected by a succinatediester linker. As a nonionic surfactant, TPGS has drawn specialattention in the drug industry as an emulsifier, stabilizer, andabsorption/permeation enhancer, and it is also a drugsolubilizer approved by the FDA.2,3 It was reported that

TPGS is readily absorbed in the gastrointestinal tracts and canincrease the oral bioavailability of some anticancer drugs byinhibiting P-glycoprotein.4 Another special interest of TPGS indrug delivery is to form TPGS based nanoparticles,2 micelles,or liposomes.5

The Fourier-transform ion cyclotron resonance (FTICR)mass spectrometer is recognized as the highest performancemass spectrometer, with high mass accuracy <1 ppm routinelyat 12 T and ultrahigh resolution. A less appreciated feature ofFTICR mass spectrometry (MS) is the flexibility provided bythe availability of various fragmentation methods. Interest inusing high performance mass spectrometry for the study ofsynthetic polymers has grown sharply in the past decadefollowing advances in polymerization methods. Especially withincreasing complexity and diversity of structures, manypolymeric materials are now requiring advanced structuralinvestigations. A high performance mass spectrometrycombined with tandem mass spectrometry (MS/MS) techni-que is a valuable tool in the study of polymer structures becauseit has been successfully used in obtaining structural information

Received: October 4, 2013Accepted: January 8, 2014Published: January 8, 2014

Figure 1. (a) Structure of TPGS. (b) Structure of Di-TPGS.

Article

pubs.acs.org/ac

© 2014 American Chemical Society 1567 dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−1574

for many categories of compounds.6−11 Study of polymersusing high performance mass spectrometry to determine theirmass distribution and absolute molar masses,12 end-groups andpolymerization mechanisms,13 structure and compositions14,15

are common areas of research. Combining LC/MS and FTICRMS/MS was employed in the end-group study of methacrylicpolymers, where the remarkable mass accuracy and resolutionof a FTICR MS instrument demonstrated its utility in studyingcomplex synthetic polymer systems.13 By using FTICR MSwith both electron capture dissociation (ECD) and collisionallyactivated dissociation (CAD), Kaczorowska and Cooperreported the structural characterization of polyphosphoesters,and their MS/MS degradation pathways were discussed.16 Mostrecently, a better understanding of the composition andvariability of two commonly used polymeric excipients,Gelucire 44/14 and polysorbate 80, between batches wereachieved aided by the high mass accuracy of FTICR MS.17

However, tandem mass spectrometry employed in thestructural characterization of synthetic polymers is stillunderdeveloped, due to difficulties involved in fragmentingdiverse, and often stable, polymer bonds between monomerunits14 and distinct data analysis strategies associated withdifferent polymer classes.TPGS is readily available from a wide range of suppliers and

manufactures. In the literature, in addition to TPGS, severalcomponents including free PEG, Di-TPGS, and α-tocopherylare reported to be present in some TPGS samples.18−20 Themain techniques currently used for the study of TPGS areNMR, FTIR, UV, and chromatography,19,21,22 but most ofthese methods struggle to provide an overview of componentcomposition of a mixture, and are particularly challenging forstructural characterization of low abundant components.Combining high performance liquid chromatography (HPLC)with mass spectrometry can provide some compositioninformation; however, most of the literature has only focusedon the information extracted from molecular mass values,without doing specific analysis of structural details using

tandem mass spectrometry due to the limited performance ofthe mass spectrometer.Herein, a high performance method using a FTICR mass

spectrometer to study the composition and structure of TPGSwas established, which could be used as a complementary toolin addition to those traditional techniques popular in polymerstudy. Utilizing FTICR MS with CAD and ECD/electroninduced dissociation (EID), MS/MS behaviors of TPGS underdifferent fragmentation methods and adducts, particularly theinfluence of different metallic cations (Li, Na, Ag), arediscussed.

■ EXPERIMENTAL METHODOLOGY

Chemicals. TPGS samples (ISOCHEM, Vert-Le-Petit,France) provided by AstraZeneca were dissolved in 50:50methanol/water or 100% acetonitrile in a concentration of ∼1μM before use. For different experiments, sodium sulfate,lithium chloride, or silver nitrate, (Fisher Scientific UK limited,UK) were added to a final concentration of 1 mM or 0.1% forformic acid (Sigma-Aldrich Co., St. Louis, MO, USA). Waterwas purified by a Millipore Direct-Q purification system(Merck Millipore, MA, USA). All other solvents werepurchased from VWR (VWR Co., Radnor, PA, USA).

Mass Spectrometry Experiments. All mass spectrometryexperiments were carried out on a Bruker 12T solariX FTICRmass spectrometer (Bruker Daltonik, GmbH, Bremen,Germany) with an electrospray source and a homemadenanoelectrospray ion source. For nanoelectrospray experi-ments, ∼5−10 μL of sample was loaded into a glass capillarytip with a stainless steel wire inside to form the electricalconnection,23 and the ionization was generated by applying avoltage of 600−1200 V between the spraying needle and inletcapillary tip. For CAD, the parent ions were isolated in thequadrupole and then transferred to the collision cell forfragmentation using a collision energy which was optimized foreach sample to provide an even spread of fragments, andfragments were finally detected in the ICR infinity cell.24 In the

Figure 2. Full mass overview of TPGS sample. Detailed peak assignment is available in the Supporting Information, Table S-1.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741568

ECD/EID experiment, the isolated ions were accumulatedexternally in the collision cell and then transferred to the ICRcell. Trapped ions were then irradiated with electrons from a1.5 A heated hollow cathode, and the bias voltages used forECD and EID experiments were ∼2 and 12 V, respectively. Forthe ECD/ECD MS3 experiment, fragment ions of interest fromECD MS/MS were isolated in the ICR cell, and then activatedwith another ECD pulse. To achieve a desirable S/N andquality spectrum, up to 200 scans were averaged on allexperiments. All spectra were phased to absorption mode25,26

and processed by DataAnalysis 4.0 software (Bruker Daltonik,GmbH, Bremen, Germany).

■ RESULTS AND DISCUSSION

Full MS Overview. For the initial study, the massdistribution of the TPGS sample was investigated by nano-electrospray FTICR MS, and the spectrum is shown in Figure2. Several groups with different charge states are present. Withthe help of high mass accuracy, the formulas of these groups areassigned to free PEG and TPGS with different adductcombinations (Na+, NH4

+, H+) and charge states. Proposedformulas are summarized in the Supporting Information, TableS-1. An overall mass accuracy around 0.4 ppm is achieved, andthe mass accuracy is generally better for peaks with higher S/Nand lower m/z. The high resolution data available from theFTICR MS is particularly useful for assigning some of theclosely spaced species. For example, in the case of the expandedinset in Figure 2, these two groups of peaks (∼20 mDa apartseparated by a resolving power of 370 000) are assigned toTPGS with different adducts, [M + 2Na + H]3+, vs [M + 3H]3

with different PEG units.As all of these formulas are proposed solely based on

experimental m/z value and restricted by the elementalcomposition of TPGS, ECD MS/MS was implemented with

several preassigned groups, [M + 3Na]3+, [M + 2Na + H]3+, [M+ 2Na + NH4]

3+, and [M + 3H]3+, to check the assignments.For the sodiated ions, after analyzing the ECD spectra of [M

+ 2Na + H]3+, [M + 2Na + NH4]3+, and [M + 3Na]3+, the

proposed formulas are consistent with the ECD results, wherepeaks losing H/NH4/Na from the molecular ions, respectively,are detected (Supporting Information, Figure S-1). Interest-ingly, the [M + 2Na + H]3+ and [M + 2Na + NH4]

3+ precursorions readily lose H and NH4, respectively, from the charge-reduced radical cation to form [M + 2Na]2+. The same twoprecursors can lose Na to form [M + Na + H]2+ and [M + Na +NH4]

2+, but this fragmentation channel is less favorable, sothese fragments are not abundant (Supporting Information,Figure S-1). Furthermore, almost all fragments from all threeprecursors are sodiated species. For alkali metal adducts,previous research using CAD showed that smaller adducts, suchas Na+, were more strongly bond to PEG backbone than biggeradducts, such as K+ or Cs+.22 In these ECD spectra, theincreased loss of H+/ NH4

+ is more favorable compared to lossof Na+, but the previous research did not do the comparison forCAD, and additionally related research is ongoing. Looking intothe fragments, most of them are from the same cleavage butwith different numbers of PEG units, and the three precursorshave very similar cleavage patterns. The interpretation of ECDfragments of [M + 3Na]3+ will be discussed in detail in the nextsection.In contrast to the sodiated precursors discussed above, the

ECD result of [M + 3H]3+ shows a difference. On the spectrumof [M + 3H]3+, a peak losing a proton or hydrogen only is notobserved, but α-tocopheryl ([C29H49O2H]

+ at m/z 430.380 53)is present as the most abundant fragment under ECD. (seeFigure S-2 in the Supporting Information) Comparing theresults, the sodium cation is more likely to attach to the PEGgroup, whereas the proton seems to stay in the tocopherylregion.

Figure 3. Full mass overview of TPGS sample in ACN with Na2SO4. Detailed peak assignment is available in the Supporting Information, Table S-2.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741569

Intriguingly, in practice, it was found that the detection of Di-TPGS (structure shown in Figure 1b) can be greatly influencedby the solution used for ionization. As discussed above, onlyfree PEG and TPGS appeared in the TPGS sample analyzedusing methanol and water (50:50 V:V) as the solution, which isa routine solution used for electrospray. However, when thesolution is switched to 100% methanol or pure acetonitrile, inaddition to the change of charge distributions of each series, Di-TPGS is also detected. Figure 3 is the mass spectrum of theTPGS sample generated by using acetonitrile with saturatedNa2SO4 as the solution. Aided by the peak assignment of Figure2 and consistent high mass accuracy (see Table S-2 in theSupporting Information), the peaks in Figure 3 are easilydefined as PEG, TPGS, and Di-TPGS with different adductsand charge states, where the corresponding peak list can befound in the Supporting Information, Table S-2. Even thoughthere are still some NH4

+ and proton adducts, most of theabundant series in Figure 3 are sodiated. Actually, the ratio oforganic solvent to water in the solution was also evaluated andpure methanol or acetonitrile was demonstrated to be the mostsuitable solution in the composition study of TPGS samples. Inaddition to a better solubility in pure organic solvent of Di-TPGS, using different percentage of organic solvent might alsoaffect the ionization mechanism of Di-TPGS during spray.Comparing the structures of TPGS and Di-TPGS in thesolution containing water, Di-TPGS could be much morefolded than TPGS, because it has two long hydrophobic tails onboth sides of its structure. Thus, during electrospray/nano-electrospray under this condition, whereas the chain ejectionmodel (CEM) may be a more accurate model for TPGS ionformation, charged residue model (CRM) appears to be moresuitable for Di-TPGS.27 Therefore, the lower CRM ionformation rate could inhibit the signal intensity of the foldedDi-TPGS. On the other hand, once the solution is switched topure organic solvent, the flexible Di-TPGS would have thesame ionization mechanism as TPGS become detectable.

Tandem Mass Spectrometry (MS/MS) of [TPGS +nNa]n+. For pharmaceutical applications, in addition toconsistent composition between batches, a consistent structureis also essential. Use of FTICR MS high quality data todetermine the composition of TPGS is discussed above. In thissection, both CAD and electron based fragmentationtechniques (ECD/EID) are applied to investigate the fragmentsof TPGS. As sodium is a ubiquitous adduct for a PEGylatedpolymer, [TPGS + 3Na]3+, [TPGS + 2Na]2+, and [TPGS +Na]+ are selected respectively in this study. TPGS with 27 PEGunits, having a formula of C87H162O32Nan (n = 1, 2, 3) isselected for further fragmentation, based on peak intensity.CAD spectra of TPGS with different numbers of sodium

adducts, [M + 3Na]3+, [M + 2Na]2+, and [M + Na]+, arepresented in Figure 4. The first interesting result is that nofragment ions in the m/z range of 98−2000 are detected for thesingly sodiated precursor, [M + Na]+, even at higherfragmentation collision energy. It is worth noting that thefragmentation of such species would be difficult using MALDIMS/MS because only singly charged ions would be generated;however, MALDI ionization is still the most commonly usedionization method for polymer study where the singly chargedions provide for a simple mass distribution.28 Fortunately, with2 or 3 sodium adducts, many cleavages are produced. (Figure4b,c) Utilizing accurate mass, the elemental formulas of mostfragments can be assigned, and peak assignments with anaverage of 0.3 ppm mass uncertainty are available in theSupporting Information, Table S-3. For sodiated species, loss ofthe entire tocopheryl succinate ester (position “f”) and a cross-ring cleavage (position “a”) are more common. The cross-ringcleavage in position “a” is intriguing, particularly as it is also themost abundant fragment in both [M + 3Na]3+ and [M +2Na]2+ CAD spectra. The intensity of this peak is likely drivenby the thermodynamic stability of a larger conjugated structureas in Scheme 1. A similar situation was reported in the preciouswork with Chlorophyll a,11 where a frequent 2H abstractionleading to a larger conjugated structure was observed.

Figure 4. CAD results of [TPGS + nNa]n+ (C87H162O32Nan) n = 1, 2, 3, and the collision energies used are the following: (a) 50, (b) 46, and (c) 35V. Detailed peak assignment is available in the Supporting Information, Table S-3.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741570

In the ECD and EID results, similarly, no fragment ions weredetected on the EID spectrum of [M + Na]+, and morefragments are produced from the higher charge-stateprecursors. (Figure 5) By utilizing a relatively higher energy(above 10 eV), EID is able to induce electron based radical ionchemistry on singly charged species, and has demonstrated itscomplementary role in obtaining structural information ofmany molecules.6,29−31 However, to our knowledge, there arefew literature reports about using EID to investigate syntheticpolymers.32 Unfortunately, in this case, EID did not generatefragments. For higher charged precursors, more charge carriersand the Coulombic repulsion between the charges apparentlycontribute to the abundant fragment ions in the MS/MSspectra. Even though it generates fewer fragments than CAD,ECD does show different bond cleavage patterns. In contrast tocleavages “a” and “f” being dominated in CAD, cleavages nextto the two carbonyl groups, at positions “d” and “e”, producethe most intense fragments in ECD of [M + 3Na]3+ and [M +2Na]2+. In terms of structural identification of TPGS, cleavagesfrom positions “a”, “f”, “d”, and “e” from CAD and ECD,provide useful, intense, diagnostic fragments.

ECD/ECD MS3. For the ECD spectrum of [M + 3Na]3+

(Figure 5c), in addition to losing one sodium and two sodiumcations, the charge-reduced species, [M + 3Na]2+•, is detected.Regarding of data analysis, apart from the restrictions of theformula of TPGS, the balance between elemental compositions,and the degree of unsaturation, one problem that makes theinterpretation of these ECD results more complicated is that asingly charged fragment can have either 1, 2, or 3 sodiumadducts, and a doubly charged ion may have 2 or 3 sodiumadducts, because of charge reduction by electron capture. Thehigh mass accuracy is of great benefit in most situations, buttandem mass spectrometry is still useful for distinguishing somesubtle features. Take the most abundant fragment, the doublycharged ion at m/z 643.339 35, as an example, the two mostpossible formulas are [C57H108O28Na2]

2+ (1.9 ppm) and[C55H109O28Na3]

2+ (0.04 ppm). If the doubly charged fragmenthas two sodium adducts, the formula may indicate a cleavagefrom the aromatic ring in the α-tocopheryl part, with a loss oftwo PEG units. However, as the average mass accuracy for theECD result is around 0.5 ppm (Table S-4 in the SupportingInformation), the formula with three sodium adducts seemsmore reasonable, which implies a cleavage from position “d”and with a loss of one PEG unit (C2H4O). To check thisassignment, ECD/ECD MS3 is employed on the fragment ionat m/z 643.339 35. Following ECD MS/MS of [TPGS +3Na]3+, all other peaks except the one of interest, m/z 643.33935, are ejected from the cell, and then another ECD pulse isapplied to the remaining ions. ECD/ECD MS3 process isshown in Figure 6. Several fragments are produced in Figure 6c,and the most useful one is the peak at m/z 1229.646 15 with aformula of C52H104O27Na3 (−0.72 ppm), which shows acleavage from position “e” and with a loss of one PEG unit. TheMS3 result clearly proves that the doubly charged peak at m/z643.339 35 is from a cleavage at position “d”, with a loss of onePEG unit. Further scrutinizing of the ECD spectrum of [M +

Scheme 1. Proposed Structure of the Fragment fromCleavage at Position “a” in CAD

Figure 5. (a) EID spectrum of [TPGS + Na]+; (b) and (c) are ECD spectra of [TPGS + 2Na]2+ and [TPGS + 3Na]3+, respectively. Detailed peakassignment is available in the Supporting Information, Table S-4.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741571

3Na]3+, revealed a peak at m/z 665.352 39 [C57H113O29Na3]3+

(0.13 ppm) representing a direct cleavage from position “d”without losing any PEG unit, which has much lower intensitythan the peak at m/z 643.339 35 (Figure S-3 in the SupportingInformation). Interestingly, in the ECD spectrum of TPGSwith two sodium adducts, [TPGS + 2Na]2+, the peak at m/z1307.715 59 [C57H113O29Na2]

+, which is from a direct cleavageat position “d” without losing any PEG unit is the mostabundant fragment. Losing one PEG unit is favorable in theECD process of [TPGS + 3Na]3+ but not for [TPGS + 2Na]2+,which is likely because of different structural conformations ofTPGS with different numbers of sodium adducts. A loss of onePEG unit (C2H4O) from the molecular ion of [TPGS + 3Na]3+

is detected, though the intensity is low (Figure S-4 in theSupporting Information).MS2 with Different Metallic Cation Adducts. It was

reported that different ionization cations can cause comple-mentary fragmentation patterns for multiply charged poly-mers,14 so the influence of several metallic-cationizing agentsare investigated. Apart from sodium, lithium and silver arechosen for this study, because lithium is the most reactive metaland the smallest alkali metal, whereas silver typically bonds to aπ bonded region. Ag+ is also known to be of beneficial forionizing and fragmenting polystyrene because of a possiblebinding to phenyl rings,33,34 which might be applicable for thechromanol ring in the tocopheryl region. Additionally, it wasreported that Li and Ag cationization are facile methods forgenerating low energy CAD spectra of polyglycols.35

CAD and ECD results of [M + 3Na]3+, [M + 3Li]3+, and [M+ 3Ag]3+ are compared in Scheme 2. TPGS with lithium cationadducts produces very similar cleavages to the one with sodiumadduct in both CAD and ECD. Comparatively, TPGS withsilver adducts also yields some similar cleavages in the CADexperiment, with more fragments in the tocopheryl side(positions “i” and “j”). Intriguingly, in the ECD spectrum of[TPGS + 3Ag]3+, no cleavages from the TPGS are generatedexcept for loss of the Ag+ cation (see Figure S-5 in theSupporting Information), though Ag has higher electron affinitythan Li or Na. A possible explanation for the loss of Ag+ cation

from the precursor is that the interaction between the PEGregion and the silver adduct is weaker than with lithium andsodium adducts, thus, silver adducts in the PEG region mightbe lost before leading to any usable fragmentation in ECD.Additionally, one of the Ag+ cations is likely to attach to thearomatic region via cation-π interaction. Because the ECD MS/MS spectrum (Supporting Information, Figure S-5) onlyinvolved partial neutralization and loss of Ag+ cation, captureof the electron to neutralize the silver atom followed by loss ofthe resulting Ag• adduct is a likely pathway. In addition,because Na+ is a common contaminant for polymers, in termsof structural characterization in practice, diagnostic peaks fromCAD and ECD of [TPGS + 3Na]3+/ [TPGS + 2Na]2+ arepreferred for simplicity.

Figure 6. (a) ECD spectrum of [M + 3Na]3+ as Figure 5c; (b) In cell isolation of the doubly charged fragment at m/z 643.3394; (c) ECD/ECD MS3

spectrum of isolated peaks. Detailed peak assignment is available in the Supporting Information, Table S-5.

Scheme 2. CAD and ECD Results of [TPGS + 3Na]3+,[TPGS + 3Li]3+, and [TPGS + 3Ag]3+

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741572

Variation of TPGS Samples. Following the compositionaland structural study using FTICR MS and MS/MS, the varietyof four batches of TPGS samples from 2 differentmanufacturers (A and B) is investigated, where the fourbatches are labeled as A1, A2, B1, and B2, respectively.According to the mass spectrum of each batch (see Figure S-6in the Supporting Information), the number average molecularweight (Mn), weight average molecular weight (Mw), repeatunit (n), and the relative intensity of TPGS, Di-TPGS and PEG(TPGS%, Di-TPGS %, and PEG%) can be calculated and aresummarized in Table 1. Table 1 shows that the polymerizationextent of TPGS for the four samples are similar, with a repeatunit of 22 ± 1 (n). Moreover, the CAD and ECD spectra of[TPGS27 + 2Na]2+ from the four samples (SupportingInformation, Figures S-7 and S-8) are almost identical, whichargues for similar structures of TPGS between batches. Eventhough consistent structures of TPGS and similar compositions(TPGS, PEG, and Di-TPGS) between the four batches areobserved, in comparison to batches B1 and B2, A1 and A2contain more TPGS and less PEG based on the relativeintensity of each series.

■ CONCLUSIONS

FTICR MS and MS/MS were demonstrated as a valuable toolin the compositional study and structural characterization ofTPGS samples, which can be used as a complementary tool fortraditional spectroscopic techniques in TPGS research. Theresults show that high mass accuracy and high resolution are ofgreat benefit in elemental formulas assignment and in obtainingdetail structural information. Some diagnostic fragments fromCAD and ECD MS/MS of [TPGS + 3Na]3+ and [TPGS +2Na]2+ can be used for the identification of the TPGS structurein the future. Interestingly, no fragment ions were detectedfrom the [TPGS + Na]+ precursor under both CAD and EID.By utilizing ECD/ECD MS3, more confident supplementarystructural information can be achieved. Varying solutionconditions were shown to affect the compositional study ofTPGS samples, indicating that the solvent selection could beparticularly important for the mass spectrometry study of somepolymers. Additionally, different metallic-cation adducts canstrongly influence the fragmentation pattern of TPGS, with thesilver adduct providing particularly interesting results. Thismethod can be applied for the rapid compositional andstructural investigation of many other polymeric compounds.

■ ASSOCIATED CONTENT

*S Supporting InformationFigure S1−8 and Table S1−6 as noted in the text. This materialis available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*P. B. O’Connor. Tel: +44 02476 151008. Fax: +44 02476151009. E-mail: p.oconnor@warwick.ac.uk.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the Chancellor’s InternationalScholarships of Warwick, Warwick Centre for AnalyticalScience (EPSRC funded EP/F034210/1), the University ofWarwick, and the Department of Chemistry. Special thanks aredue to Mark P. Barrow, Huilin Li, Pilar Perez-Hurtado, RebeccaWills, Yulin Qi, Andrea Lopez-Clavijo, Andrew Soulby,Samantha Benson, Maria van Agthoven, Chris Wootton fromUniversity of Warwick, and Andrew Ray from AstraZeneca.

■ REFERENCES(1) Yan, A.; Von Dem Bussche, A.; Kane, A. B.; Hurt, R. H. Carbon2007, 45, 2463−2470.(2) Mu, L.; Seow, P. H. Colloids Surf., B 2006, 47, 90−97.(3) Guo, Y.; Luo, J.; Tan, S.; Otieno, B. O.; Zhang, Z. Eur. J. Pharm.Sci. 2013, 49, 175−186.(4) Dintaman, J.; Silverman, J. Pharm. Res. 1999, 16, 1550−1556.(5) Zhang, Z.; Tan, S.; Feng, S.-S. Biomaterials 2012, 33, 4889−4906.(6) Wills, R. H.; Tosin, M.; O’Connor, P. B. Anal. Chem. 2012, 84,8863−8870.(7) Li, H.; Lin, T.-Y.; Van Orden, S. L.; Zhao, Y.; Barrow, M. P.;Pizarro, A. M.; Qi, Y.; Sadler, P. J.; O’Connor, P. B. Anal. Chem. 2011,83, 9507−9515.(8) Yu, X.; Huang, Y.; Lin, C.; Costello, C. E. Anal. Chem. 2012, 84,7487−7494.(9) Jin, Z.; Daiya, S.; Kenttamaa, H. I. Int. J. Mass spectrom. 2011, 301,234−239.(10) Taucher, M.; Breuker, K. Angew. Chem. 2012, 124, 11451−11454.(11) Wei, J.; Li, H.; Barrow, M.; O’Connor, P. J. Am. Soc. MassSpectrom. 2013, 24, 753−760.(12) O’Connor, P. B.; McLafferty, F. W. J. Am. Chem. Soc. 1995, 117,12826−12831.(13) Song, J.; van Velde, J. W.; Vertommen, L. L. T.; Smith, D. F.;Heeren, R. M. A.; van den Brink, O. F. Macromolecules 2011, 44,1319−1326.(14) Miladinovic, S. M.; Kaeser, C. J.; Knust, M. M.; Wilkins, C. L.Int. J. Mass Spectrom. 2011, 301, 184−194.(15) Cerda, B.; Horn, D.; Breuker, K.; Carpenter, B.; McLafferty, F.Eur. J. Mass Spectrom. 1999, 5, 335−338.(16) Kaczorowska, M. A.; Cooper, H. J. J. Am. Soc. Mass Spectrom.2009, 20, 2238−2247.(17) Perez Hurtado, P.; Lam, P. Y.; Kilgour, D.; Bristow, A.;McBride, E.; O’Connor, P. B. Anal. Chem. 2012, 84, 8579−8586.(18) Kong, L. Y.; Su, B. G.; Bao, Z. B.; Xing, H. B.; Yang, Y. W.; Ren,Q. L. J. Chromatogr. A 2011, 1218, 8664−8671.(19) Wang, J.; Su, B.; Xing, H.; Yang, Y.; Ren, Q. J. Appl. Polym. Sci.2011, 119, 3026−3033.

Table 1. Comparison of Four Batches of TPGS Samples from Two Manufacturers

TPGS Di-TPGS PEG

batch Mna Mwb nc Mn Mw n Mn Mw n TPGS % Di-TPGS % PEG%

A1 1461.751 1484.335 21.2 2039.141 2051.342 22.6 905.9376 924.6729 20.2 90.8 2.7 6.4A2 1505.268 1529.107 22.1 1951.339 1955.496 20.6 955.8597 981.0104 21.3 89.4 3.0 7.6B1 1502.263 1524.667 22.1 2046.761 2057.954 22.8 953.2361 977.8347 21.2 81.4 4.9 13.7B2 1520.241 1544.232 22.5 2094.985 2105.985 23.9 957.8828 978.756 21.3 84.9 2.8 12.2

aMn: Number average molecular weight. bMw: Weight average molecular weight. cn: Repeat unit.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741573

(20) Christiansen, A.; Backensfeld, T.; Kuhn, S.; Weitschies, W. J.Pharm. Sci. 2011, 100, 1773−1782.(21) Cao, N.; Feng, S.-S. Biomaterials 2008, 29, 3856−3865.(22) Bogan, M. J.; Agnes, G. R. J. Am. Soc. Mass Spectrom. 2002, 13,177−186.(23) Zhao, C.; Sethuraman, M.; Clavreul, N.; Kaur, P.; Cohen, R. A.;O’Connor, P. B. Anal. Chem. 2006, 78, 5134−5142.(24) Caravatti, P.; Allemann, M. Org. Mass Spectrom. 1991, 26, 514−518.(25) Kilgour, D. P.; Wills, R.; Qi, Y.; O’Connor, P. B. Anal. Chem.2013, 85, 3903−3911.(26) Qi, Y.; Barrow, M. P.; Li, H.; Meier, J. E.; Van Orden, S. L.;Thompson, C. J.; O’Connor, P. B. Anal. Chem. 2012, 84, 2923−2929.(27) Konermann, L.; Ahadi, E.; Rodriguez, A. D.; Vahidi, S. Anal.Chem. 2012, 85, 2−9.(28) Mautjana, N. A.; Pasch, H. Macromol. Symp. 2012, 313−314,157−161.(29) Lioe, H.; O’Hair, R. Anal. Bioanal. Chem. 2007, 389, 1429−1437.(30) Mosely, J. A.; Smith, M. J. P.; Prakash, A. S.; Sims, M.; Bristow,A. W. T. Anal. Chem. 2011, 83, 4068−4075.(31) Wolff, J. J.; Laremore, T. N.; Aslam, H.; Linhardt, R. J.; Amster,I. J. J. Am. Soc. Mass Spectrom. 2008, 19, 1449−1458.(32) Smith, M. J. P., Ph.D. Dissertation, Department of Chemistry,Durham University, U.K., 2013.(33) Jackson, A. T.; Williams, J. P.; Scrivens, J. H. Rapid Commun.Mass Spectrom. 2006, 20, 2717−2727.(34) Deery, M. J.; Jennings, K. R.; Jasieczek, C. B.; Haddleton, D. M.;Jackson, A. T.; Yates, H. T.; Scrivens, J. H. Rapid Commun. MassSpectrom. 1997, 11, 57−62.(35) Chen, R.; Li, L. J. Am. Soc. Mass Spectrom. 2001, 12, 832−839.

Analytical Chemistry Article

dx.doi.org/10.1021/ac403195f | Anal. Chem. 2014, 86, 1567−15741574