Synthetic oligomer analysis using atmospheric pressure photoionization mass spectrometry at different photon energies

Analytica Chimica Acta 808 (2014) 220–230

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Synthetic oligomer analysis using atmospheric pressure photoionization mass spectrometry at different photon energies Bernard Desmazières b , Véronique Legros a , Alexandre Giuliani c,d , William Buchmann a,∗ a

CNRS, UMR8587, Université d’Evry-Val-d’Essonne, Laboratoire Analyse et Modélisation pour la Biologie et l’Environnement, F-91025 Evry, France Global Bioenergies, 5 rue Henri Desbruyeres, 91030 Evry, France Synchrotron SOLEIL, L’Orme des Merisiers, Saint-Aubin, 91192 Gif-sur-Yvette, France d UAR1008, CEPIA, INRA, Rue de la Geraudiere, F-44316 Nantes, France b c

h i g h l i g h t s

g r a p h i c a l

• Atmospheric

Atmospheric pressure photoIonization mass spectra of synthetic oligomers were recorded in the negative mode by varying the photon energy using synchrotron radiation. Photon energy required for an efficient ionization of the polymer was correlated to ionization potential of the solvent (for example 9.4 eV for tetrahydrofuran).

• • • •

pressure photoionization was performed using synchrotron radiation. Photoionization of oligomers in THF with 10% CH2 Cl2 produces intact [M + Cl]− ions. The photon energy required corresponds to ionization potential of the solvent. Polymer distributions depend on source parameters such T ◦ C and applied voltages. Liquid chromatography was coupled to MS using an APPI interface for polymer analysis.

a r t i c l e

i n f o

Article history: Received 8 July 2013 Received in revised form 15 November 2013 Accepted 16 November 2013 Available online 22 November 2013 Keywords: Photoionization Polymer Mass spectrometry Chromatography Synchrotron light Atmospheric pressure photoionization

a b s t r a c t

a b s t r a c t Atmospheric pressure photoionization (APPI) followed by mass spectrometric detection was used to ionize a variety of polymers: polyethylene glycol, polymethyl methacrylate, polystyrene, and polysiloxane. In most cases, whatever the polymer or the solvent used (dichloromethane, tetrahydrofuran, hexane, acetone or toluene), only negative ion mode produced intact ions such as chlorinated adducts, with no or few fragmentations, in contrast to the positive ion mode that frequently led to important in-source fragmentations. In addition, it was shown that optimal detection of polymer distributions require a fine tuning of other source parameters such as temperature and ion transfer voltage. Series of mass spectra were recorded in the negative mode, in various solvents (dichloromethane, tetrahydrofuran, hexane, toluene, and acetone), by varying the photon energy from 8 eV up to 10.6 eV using synchrotron radiation. To these solvents, addition of a classical APPI dopant (toluene or acetone) was not necessary. Courtesy of the synchrotron radiation, it was demonstrated that the photon energy required for an efficient ionization of the polymer was correlated to the ionization energy of the solvent. As commercial APPI sources typically use krypton lamps with energy fixed at 10 eV and 10.6 eV, the study of the ionization of polymers over a wavelength range allowed to confirm and refine the previously proposed ionization mechanisms. Moreover, the APPI source can efficiently be used as an interface between size exclusion chromatography or reverse phase liquid chromatography and MS for the study of synthetic oligomers. However, the photoionization at fixed wavelength of polymer standards with different molecular weights showed that it was difficult to obtain intact ionized oligomers with molecular weights above a few thousands. © 2013 Elsevier B.V. All rights reserved.

∗ Corresponding author. Tel.: +33169477646; fax: +33169477655. E-mail address: [email protected] (W. Buchmann). 0003-2670/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2013.11.036

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

1. Introduction Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) is currently the most popular MS technique for polymer analysis [1,2]. This technique allows characterization of repeat units and end-groups of synthetic polymers. From a MALDI mass spectrum, reliable values of number average molecular weight (Mn ), weight average molecular weight (Mw ) and of polydispersity index of a molecular weights distribution (PDI, defined by Mw /Mn ) can be deduced for narrow polydispersity polymers (PDI <1.2) [3,4]. Tandem MS experiments with MALDI-TOF–TOF instruments bring an additional dimension in structural characterization of polymers [5]. Nowadays, MALDI-MS is the most widespread MS technique for polymer analysis, but alternative ionization techniques are desirable for direct coupling with liquid chromatography (LC) or for compounds that are difficult to be ionized by MALDI. Polymer analyses using ambient ionization methods such as desorptionelectrospray ionization (DESI) [6–11], direct analysis in real time (DART) [12] and atmospheric solids analysis probe (ASAP) [13] techniques, were recently published, but these latter techniques cannot be directly used as interfaces with liquid chromatography (LC). Among the commonly used techniques as LC–MS interfaces, such as electrospray ionization (ESI) [14,15], atmospheric pressure chemical ionization (APCI) [16,17] and atmospheric pressure photoionization (APPI) [18,19], the last appears to be as a promising ionization technique for MS and LC–MS analyses of synthetic polymers, especially for nonpolar compounds, despite the small number of articles recently published [20–23]. Only three articles from Zsuga et al., deal with the use of APPI for polymer analysis [20–22]. Low mass polyethylene and polyisobutylene with different endgroups were studied using a commercial APPI source (Photomate) equipped with a Kr discharge lamp emitting vacuum ultraviolet (VUV) photons of 10.0 eV and 10.6 eV in an intensity ratio of 4:1. This APPI source was combined with a hybrid Q-TOF mass analyzer. In the positive ion mode, significant fragmentation was noted by Zsuga et al. whereas in the negative mode, abundant chlorinated adduct ions were produced from chlorinated solvents such as carbon tetrachloride in the presence of toluene as dopant [20–22]. The reported results were all in the flow injection mode only, and at fixed wavelength. In the present work, the potential of the vacuum ultraviolet (VUV) photons for the ionization of polymers was explored over a wavelength range from 8 eV up to 10.6 eV using the coupling of an APPI source with a VUV synchrotron radiation beamline (DISCO) at the SOLEIL synchrotron radiation facility in France. Using this set-up as described by Giuliani et al. [24], the ionization of guanine in the positive mode was recently studied by Touboul et al. as a function of the photon energy [25]. The use of the coupling of the synchrotron radiation with MS was also reported by Le Naour et al. for the characterization of peptide palmitoylation, a major post-translational modification of membrane proteins [26]. By studying the formation of gaseous ions by APPI-MS from various synthetic polymers (polyethylene glycol, polymethyl methacrylate, polydimethyl siloxane and polystyrene) at different wavelengths, one of our objectives was to gain a better understanding of the mechanism of formation of ions. The crucial questions for polymer analysis that are related to the sensitivity, mass range, and the kind of polymer able to be analyzed by APPI-MS were also addressed. In addition, we report here the first coupling experiments of LC with APPI-MS for synthetic polymer analysis. LC–MS coupling can be a powerful tool for polymer analysis [27]. The combination of chromatographic separation and identification by APPI-MS may reduce the complexity of the analysis of polymers. Finally, this work was completed by tandem MS experiments in

221

order to evaluate the amount of information that could be obtained by CID-MS/MS from the chlorinated adducts produced by APPI.

2. Experimental 2.1. Materials Polydimethyl siloxane (PDMS 20 cs, Mn ∼ 2000 g mol−1 ), was purchased from Sigma-Aldrich (Saint Quentin-Fallavier, France), used without further purification. Polyethylene glycol (PEG, Mp 620 g mol−1 , 1080 g mol−1 , 1470 g mol−1 and 4120 g mol−1 ), and polymethyl methacrylate (PMMA, Mp 855, 1970, 2710 g mol−1 ), and polystyrene (PS, Mp 580 g mol−1 , 1355 g mol−1 and 4950 g mol−1 ) were obtained as Gel Permeation Chromatography (GPC) standards from Polymer Laboratories (now a part of Varian, Santa Clara, CA, USA). Mp corresponds to the most probable molecular weight of the molecular distribution given by the supplier determined by GPC. All polymers were used after dissolution in solvent (10−3 M). No salt was added. All solvents were HPLC grade. Dichloromethane unstabilized, dichloromethane stabilized with amylene, toluene and tetrahydrofurane unstabilized were purchased from Fisher Scientific (Illkirch, France). Dibromomethane and acetonitrile were acquired from Sigma-Aldrich (Saint Quentin Fallavier, France). Acetone was purchased from Merk (Fontenay Sous Bois, France) and tetrahydrofurane stabilized with BHT (3,5di-tert-butyl-4-hydroxytoluene) was from Acros organics (Illkirch, France).

2.2. Instrument set-up Experiments were carried out using a hybrid Q-TOF (Q STAR Pulsar i, ABSciex) mass spectrometer equipped with a PhotosprayTM Source (APPI). The APPI source has been modified to allow the introduction of photons from the DISCO beamline [28] of a synchrotron radiation facility (SOLEIL) by substitution of the standard UV krypton lamp [24]. Photon energy was tunable from 5 eV (248 nm) to 20 eV (62 nm) with a precision of 0.1 nm. Source temperature was 300–500 ◦ C. In the negative ion mode, typical applied voltages were −100 V for focusing potential, 10 V for declustering potential (1) and −15 V for the declustering potential (2). The gas pressures (N2 ) in the APPI source were set at 40 psi for GAS 1 and 20 psi for GAS 2. The curtain gas pressure was 20 psi and that of collision gas was 3 psi. The samples were introduced using the flow-injection analysis (FIA) method: 20 ␮L of the sample solution at 10−3 mol L−1 was loaded into an injection loop and pushed through the APPI source by the solvent at a flow rate of 200 ␮L min−1 . All LC-MS experiments (RPLC and SEC) were performed using an HP1100 liquid chromatograph (Agilent Technologies). Reversedphase chromatographic separation was performed on a C18 column (Fused-Core Ascentis, 100 × 3 mm, diameter particle 2.7 ␮m) with the following gradient: linear gradient from acetonitrile (with 5% acetone) to 65% of dichloromethane, return to acetonitrile with 5% acetone within 5 min, and equilibrate the column at those conditions for 10 min, at a flow rate of 200 ␮L min−1 . SEC analyses were carried out using a PLgel column in isocratic mode (Polymers Labs, France), 300 × 8 mm, pore diameter ˚ particle diameter 5 ␮m (polydivinylbenzene/polystyrene 1000 A, copolymer). THF at a flow rate of 200 ␮L min−1 was used as eluent. The injection volume was 20 ␮L (concentration 10−3 M) for all LC and SEC-MS analyses. In mass spectra, the chemical background induced by the solvent was systemically removed by signal subtraction.

222

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

3. Results and discussion 3.1. Characteristics of APPI for polymers-influence of various parameters Fig. 1 displays a APPI mass spectrum recorded at 10.6 eV photon energy from polystyrene 1350 (Mp = 1350 g mol−1 , hydrophobic polymer) in the negative ion mode. [M + Cl]− adducts were formed in the presence of toluene as dopant, in CH2 Cl2 as solvent. Two consecutive ions are separated by 104 u as expected, corresponding to the molecular weight of the styrene repeat unit. Similarly, APPI mass spectra with intense signals were obtained from a variety of low mass polymers whatever their hydrophobicity or polarity: polyethylene glycol (PEG), polymethylmethacrylate (PMMA), polystyrene (PS) and polysiloxanes (PDMS). The quality of the results strongly depends on the chosen ionization mode (positive or negative). In most cases, whatever the solvent (CH2 Cl2 , THF, hexane or toluene), only the negative ion mode produces intact ions with no or a few fragmentations. Whatever the polymer (PEG, PS or PDMS) chlorinated adducts were systematically formed in gasphase in the negative ion mode. In contrast, the positive ion mode frequently led to undesirable phenomena of decomposition. Fragmentations in the positive mode are difficult to interpret because the decomposition takes place from a mixture of oligomers and not from a unique compound. But it can be assumed that theses fragmentations are consecutive to the production of [M + cation]+ , [M + H]+ or M•+ . In order to limit the undesirable fragmentations of polymers, frequent in the positive mode, our Q-TOF mass spectrometer was slightly modified by installing an ion cooler guide in the interface region [29]. We initially thought that the fragmentation of ions during the APPI process could come from collision-induced dissociation (CID) in the interface region of the Q-TOF mass spectrometer (up-front CID) and that poor ionization of the heavier oligomers might also be due to these unwanted fragmentations. Thus, it was expected that collisional cooling of the ions would lead to a better detection of the higher molecular weights. Unfortunately, collisional cooling in the interface region did not provide any benefits, irrespective of the ion mode positive or negative, the

solvent used, and the polymer. Fragmentation of polymers was still observed in the positive ion mode and the detection of intact heavier oligomers was not improved (data not shown). Even if we did not obtain the expected results, one can now conclude that fragmentations take place in the source region from protonated or cationized species, or radical-cations, rather than in the interface region. Since the negative mode led to simpler results than the positive mode, the negative mode was finally preferred. Consequently, only mass spectra recorded in the negative mode will be presented in this article. Most solvents contain a sufficient amount of chlorinated species to allow chlorinated adduct formation in the negative mode, but by adding 5% to 10% of CH2 Cl2 , the source of chlorine is perfectly controlled. In THF, different proportions of CH2 Cl2 were tried: 1%, 5%, 10% and 25%. 5–10% provided the best results in terms of sensitivity. It is important to note that the addition of 10% CH2 Cl2 to the solvent does not modify the position of the main energy threshold required for ionization (detailed below). Other halogenated solvents (CH2 Br2 , CHCl3 and CCl4 ) were tried instead of CH2 Cl2 , but they did not provide better signal over noise ratios. Addition of CH2 Br2 led to abundant brominated adducts [M + Br]− . An example of brominated adducts distribution from PS 1350 is shown in Fig. 2a. The Br atom possesses two isotopes similarly to the Cl atom, but with different masses and relative abundances (79 Br/81 Br 1/1; 35 Cl/37 Cl 3/1). Thus, there is a shift of 44 u between the two distributions of molar masses Fig. 2a and b and isotopic profiles are not identical (note that in Fig. 2a, for n = 14, monoisotopic ion is not the major one). THF and CH2 Cl2 are unstable compounds that degrade with time upon storage. Since THF and CH2 Cl2 are often commercialized with additives to improve their stability, a possible effect of their presence was also examined. THF can be stabilized by the manufacturer by the addition of 0.1–0.3‰ butylated hydroxytoluene (BHT, ionization potential IP 7.8 eV), CH2 Cl2 can be stabilized by the addition of ∼1% amylene (IP 8.69 eV), but by comparing results obtained from solvents stabilized and not stabilized at various photon energies, it was concluded that the influence of these stabilizers was not significant (data not shown). From the various polymers (PEG, PMMA, PS and PDMS), the highest mass ever detected was ∼6000 Da with PDMS in THF by

Fig. 1. APPI mass spectrum of polystyrene 1350 (GPC standard, Mn = 1285 g mol−1 , Mp = 1355 g mol−1 , Ip = Mw /Mn = 1.07). Negative ion mode, dopant = toluene 50 ␮L min−1 , solvent = CH2 Cl2 , h = 10.6 eV, [PS] = 10−3 M in CH2 Cl2 , T = 450 ◦ C, flow rate = 200 ␮L min−1 .

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

223

Fig. 2. (A) APPI mass spectrum of polystyrene 1350 in THF with 10% CH2 Br2 and (B), APPI mass spectrum of polystyrene 1350 in THF with 10% CH2 Cl2 (Negative ion mode, [PS] = 10−3 M, T = 500 ◦ C, flow rate = 200 ␮L min−1 ). (C) MS/MS spectrum of a brominated adduct at m/z 1594.1 from polystyrene 1350, (D) MS/MS spectrum of a chlorinated adduct at m/z 1550.2 from polystyrene 1350.

224

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

Fig. 3. APPI mass spectrum of PDMS 20 cs (Negative ion mode, without dopant, solvent: THF, white beam, 10−3 M in CH2 Cl2 , T = 450 ◦ C, flow rate = 200 ␮L min−1 ). The white beam contains all the photon energies up to 10.6 eV, which is the transparency limit of the window. In this mode, the monochromator is used as a mirror.

using a white beam (Fig. 3), but in general it was difficult to reach molecular weights above a few thousands. We tried to ionize series of GPC standards of PEG, PMMA and PS with different Mw (roughly from 500 to 5000) by APPI at fixed wavelength (data not shown). The heavier chains, if they are vaporized/ionized, undergo decomposition. Moreover, optimal detection of polymers also depends

on source parameters such as temperature, ion transfer voltage (IS) and flow rate. Fig. 4 shows the evolution of the mass spectra of PEG 1470 as a function of probe temperature. Temperature has a strong influence on the distributions of molar masses. The higher temperature allows the detection of the longer chains. Only a high temperature allowing complete desolvation of the species

Fig. 4. APPI mass spectra of PEG 1470 (GPC standard, Mn = 1444 g mol−1 , Mp = 1471 g mol−1 , Ip = Mw /Mn = 1.03) recorded as a function of probe temperature (Negative ion mode, without dopant, solvent: THF, h = 10.6 eV, 10−3 M in THF, flow rate = 200 ␮L min−1 ).

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

225

Fig. 5. APPI mass spectra of PMMA 2710 (GPC standard, Mn = 2500 g mol−1 , Mp = 2710 g mol−1 , Ip = Mw /Mn = 1.08) recorded as a function of ion transfer voltage IS (negative ion mode, without dopant, solvent = THF, h = 10.6 eV, 10−3 M in THF, T = 500 ◦ C, flow rate = 200 ␮L min−1 ).

can give an ion distribution approaching the actual distribution of the molecular weights. Fig. 5 displays signal evolution as a function of ion transfer voltage (IS). In the case of PMMA 2710, the ion distribution corresponds to a [M − CH3 ]− distribution that would come from the loss of CH3 Cl from the [M + Cl]− adducts. This loss of CH3 Cl can be minimized by lowering the temperature, but that induces a shift of the ion distribution towards the low masses. The influence of the IS voltage is less important compared to the temperature, but can affect the observed distribution. IS voltage can also be involved in the desolvation of the species. Fig. 5 shows that the lower IS voltages allow transmitting heavier molar masses distributions (−1300 V was an optimal value). From these considerations, it becomes clear that the results concerning the MW distributions deduced from APPI-MS data must be taken with care and that mass spectra recording requires a fine tuning of conditions. 3.2. Mechanism of ion formation The experimental conditions used in Fig. 1 for PS 1350 in the negative ion mode, namely with addition of toluene as dopant, at 10.6 eV energy photons can be considered as standard APPI conditions. Based on the ionization mechanisms previously proposed by Kauppila et al. [30,31] and Zsuga et al. [20,21], it can be suggested that first; photons react with toluene according to the following reaction: •

h + toluene → toluene+ + e−

(1)

if the photon energy (h) is above the ionization energy (IE) of toluene (8.83 eV [32]). It is worth noting that in the Syagen APPI source (Agilent, Thermo Scientific), the UV lamp is directed toward a metallic surface and this surface can act as a photoelectron source [33]. In our case, the geometry of the Photospray® source from ABSciex is less favourable. Thus, the contribution of photoelectrons is less likely, but one cannot exclude their presence. Then chloride

anions may be released from the solvent by dissociative electron capture according to: • − e− + CH2 Cl2 → CH2 Cl−• 2 → CH2 Cl + Cl

(2)

The last step would correspond to adduct formation by the attachment of the Cl− ion to the oligomer: Cl− + oligomer → [oligomer + Cl]−

(3)

Halide anion-attachment has also been reported for compounds other than polymers [34–36]. Under classical APPI conditions, if the solvent cannot be ionized by the VUV radiation, then addition of a dopant, which can be photoionized is necessary. However, same ions could be produced in our case without toluene addition (data not shown). Considering that IE (CH2 Cl2 ) = 11.33 eV > h = 10.6 eV, CH2 Cl2 could not be photoionized. Thus, it was first hypothesized that PS might act as a dopant itself. Surprisingly, when PS 1350 was introduced in other solvents (THF 9.4 eV, hexane 10.13 eV, toluene 8.83 eV), [32] the same chlorinated adducts were still produced. This surprising result indicates that trace amounts of chlorinated species were sufficient for the ionization success. In order to understand the sequence of events leading to formation of chlorinated adduct in the negative ionization mode without dopant addition, a great number of APPI mass spectra of various polymers (PS, PMMA, PEG and PDMS) were recorded in different solvents (THF, toluene, hexane, acetone) by varying photon energy from 8 eV up to 10.6 eV. Fig. 6a shows a typical ion current obtained from PS 1350 in THF with increasing photon energies and Fig. 6b, the mass spectrum of PS 1350 recorded at 9.4 eV (note that this mass spectrum is similar to that presented in Fig. 1, the shift of the ions distribution towards the heavier masses is mainly due to the temperature, which is 50 ◦ C higher). Fig. 7a shows the ion intensity arising from three different polymers in THF as a function of photon energy, Fig. 7b, the ion intensity coming from PS 1350 only, in three different solvents as a function of photon energy. From

226

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

Fig. 6. (A) APPI total ion current from h = 7.4 eV up to 10.6 eV of PS 1350 and (B) mass spectrum of PS 1350 (GPC standard, Mn = 1285 g mol−1 , Mp = 1355 g mol−1 , Ip = Mw /Mn = 1.065) at h = 9.4 eV (Negative ion mode, solvent: THF, [PS] = 10−3 M, T = 500 ◦ C, flow rate = 200 ␮L min−1 ).

these experiments, it was established that polymer ions appear from a particular photon energy threshold. For a given solvent, the required photon energy threshold to obtain ions is always the same whatever the polymer (Fig. 7a). For a given polymer, it was observed that the required photon energy threshold varied as a function of the solvent only (Fig. 7b). This energy threshold corresponds to the ionization energy of the solvent (8.83 eV for toluene, 9.4 eV for THF, 9.7 eV for acetone, 10.13 eV for hexane) [32]. This clearly demonstrates that the APPI process is triggered by the ionization of solvent. Below the ionization threshold of the solvents, polymer ions cannot be detected. In practice THF, toluene, hexane and acetone can efficiently be ionized using the classical krypton discharge lamp (10 eV and 10.6 eV), but neither methanol (10.84 eV) nor acetonitrile (12.2 eV) may be ionized at the wavelength of the Kr lamp. Among the various solvents tested (THF, toluene, hexane, acetone, dichloromethane, methanol, water, acetonitrile), one solvent, dichloromethane, produces polymer ions by the APPI process below its ionization threshold (11.33 eV) and the same behaviour was observed whatever the supplier of CH2 Cl2 . Curiously, first ions appear at 9.2–9.4 eV. This is probably due to the presence of an impurity of low IE in CH2 Cl2 , or to in-source photoinduced chemical reactions that produce ionizable species. Such an unexpected behaviour has already been described for acetonitrile [33,37]. Concerning the possible presence of an impurity of low IE, common reported impurities in commercial dichloromethane are hydrogen chloride (12.74 eV), phosgene (COCl2 , 11.55 eV), chloroform (11.37 eV), 1,1,2,2-tetrachloroethane (11.1 eV), methanol (10.84 eV), water (12.62 eV), and some decomposition products such as formaldehyde (10.88 eV), but the IE values of these impurities are too high [38].

Taken together, the data allows us to propose with confidence the following general ionization mechanism. First, photons react with solvent according to the following reaction: •

solvent + h → solvent+ + e− , Then according to Eq. (5), X−

if h > IE (solvent) halide anions (Cl−

(4)

or Br− ) could be

formed from halogenated species by dissociative electron capture, the attachment of X− to oligomers would occur in a last step (6). e− + CHz X4−z → CHz X−• → CHz X•3−z + X− . 4−z

(5)

X− + oligomer → [oligomer + X]− .

(6)

[oligomer + X]−

Thus, stable adducts are formed in the APPI source. Some additional remarks can be made: (i) the fact that photon energy must surpass a certain threshold to initiate the ionization enables us to exclude a simple thermospray contribution in the ionization mechanism, it was checked that photons were necessary (no photon means no ion), (ii) during the experiment consisting in the variation of the photon energy, no discrimination effect according to the Mw was noted (for example, higher energies did not lead to higher mass oligomers detection, an increase of photon energy did not lead to an increase of oligomer fragmentation). Additionally, it is worth noting that Fig. 7a displays different ion intensities increasing along with the increase of the photon energy. At 9.4 eV, the observed response differences are related to polymer nature since polymers were injected at the same concentration, with the same flow rate. One can note the following ranking of ionization yields of polymers in THF (PMMA > PS > PEG). This ranking is maintained whatever the solvent: toluene, THF, acetone (data not shown). From Fig. 7b, one can also observe that

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

227

Fig. 7. (A) Evolution of APPI ion intensity of different polymers (PS 1350, PMMA 1970, PEG 1470) in THF as a function of photon energy. (B) Evolution of ion intensity of PS 1350 in different solvents (THF, acetone, toluene) as a function of photon energy.

Fig. 8. Separation by SEC-APPI MS of a mixture of polystyrenes 580/1350/4950 g mol−1 . Total ion current and mass spectra of 3 fractions. SEC column: Polymer Labs 300 mm × 8 mm, 1000 A˚ pore diameter, 5 ␮m particle diameter (polydivinylbenzene/polystyrene), isocratic mode, negative ion mode, without dopant, solvent = THF, h = 10.6 eV, T = 500 ◦ C, flow rate = 200 ␮L min−1 , injection volume = 20 ␮L, concentration = 10−3 M in THF.

228

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

toluene provides a better ionization yield for PS for instance than THF does, and that THF provides more intense signals than acetone does. The ionization yield is clearly affected by the solvent used, but the most suitable solvent can be different for each polymer. For example, among the three solvents used (toluene, THF, acetone), toluene was the best solvent for PS and PMMA, whereas THF gave the best results for PEG. The observed differences of signal intensities are related to the efficiency of the photoionization process of not only the solvent, but they could also be modulated by affinity differences of the polymers for halides. Moreover, the ionization efficiency of the dopant molecules (the solvent in our case) depend on their molar flow rates and on the photoionization cross-section of the solvent molecule at the wavelength of the photons [39]. 3.3. SEC/APPI-MS and RPLC/APPI-MS couplings To our knowledge, the use of the LC-APPI-MS coupling has never been reported for polymer analysis [23]. Two examples of results obtained from LC couplings with APPI-MS are given

in Figs 8 and 9. In Fig. 8, a mixture of PS (580/1350/4950) was analyzed by SEC-APPI MS. SEC (or gel permeation chromatography GPC) is a low-resolution chromatographic technique, which often shows broad peaks that are not baseline separated. According to this separation mode, the oligomers with the higher molecular weights are the first to elute, then the lighter oligomers elute later. This kind of experiment allows the understanding of the behavior of the heavier chains under APPI. Despite the fact that the chromatographic peak exhibits an unusual profile, the GPC separation is truly effective. APPI detection of intact oligomers was possible for low mass oligomers between 1000 Da and 2000 Da. In contrast, the on-line MS analysis of the GPC fraction 1 (containing oligomers with M > 2000 g mol−1 ), gives a distribution of fragment ions with a decrease of detection sensitivity. This shows that for oligomers of higher mass, vaporization/ionization steps are more difficult to achieve and then fragmentations occur. The detection of intact PS 4950 is not possible. Reverse Phase LC (RPLC) separations using C18 bonded silica followed by APPI-MS detection were also carried out from PS (see the example with PS 580 in Fig. 9). In

Fig. 9. Separation by RPLC-APPI MS of a polystyrene (GPC standard, Mn = 582 g mol−1 , Mp = 572 g mol−1 , Ip = Mw /Mn = 1.18). Total ion current and mass spectra of three fractions. Column C18 Fused-Core Ascentis, 100 mm × 3 mm, 2.7 ␮m particle diameter, thickness 0.5 ␮m. Negative ion mode, without dopant. Gradient program: acetonitrile (with 5% acetone) → 20 min → 65% CH2 Cl2 (5 min) → 5 min → acetonitrile (with 5% acetone). h = white beam, T = 500 ◦ C, flow rate = 200 ␮L min−1 , injection volume = 20 ␮L, concentration = 10−2 M.

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

229

Fig. 10. (A) mass spectrum of PEG 1470 in THF with 10% CH2 Cl2 (Negative ion mode, [PS] = 10−3 M, T = 500 ◦ C, flow rate = 200 ␮L min−1 ), (B) MS/MS spectrum of a chlorinated adduct at m/z 1550.1 from PEG 1470, (C) MS/MS spectrum of a brominated adduct at m/z 1594.1 from PEG 1470 in the presence of 10% CH2 Br2 .

this case, the elution order was the opposite. The separation was good enough to fully separate different oligomers. Mass spectra with intact ions are displayed in Fig. 9. These two examples illustrate the major advantage that APPI-MS possesses with its suitability to couple chromatographic methods. The compatibility with a much larger range of solvents than ESI allows the coupling with the chromatographic modes SEC and RPLC in pure organic solvents. Clearly, APPI technique creates new opportunities for the characterization of polymers, especially for low molecular weight compounds or for the polymers that are poorly ionisable by the usual ionization methods, typically the most apolar polymers. 3.4. APPI-MS/MS experiments For each polymer (PS, PEG and PMMA), halogenated adducts were mass selected and then submitted to collision with N2 gas to

induce dissociation. A few examples are given in Figs. 2 and 10 with PS 1350 and PEG 1470. The results were disappointing because very poor MS/MS spectra were obtained. From [PS14 + X]− , the major or only fragmentation pathway was the loss of one halogenated ion Cl− or Br− (see Fig. 2C and D). From [PEG34 + X]− , the major fragmentation pathway was the loss of CH2 = CHOH, Cl− or Br− (see Fig. 10B and C). The presence of one halogenated atom in the product ions was checked by systematically mass selecting [PS14 + 35 Cl]− followed by [PS14 + 37 Cl]− , and [PEG34 + 79 Br]− followed by [PEG34 + 81 Br]− . Masses of product ions were shifted as expected. In contrast, the fragmentation of the PMMA ions produced by APPI gave some ions in the low mass region separated by 100 u (typically m/z 87, 101, 187, 201, 255), but it is significant to recall that in the case of PMMA, the distribution of ions corresponds to a [M − CH3 ]− distribution and not [M + Cl]− adducts. Finally, MS/MS experiments have essentially confirmed that

230

B. Desmazières et al. / Analytica Chimica Acta 808 (2014) 220–230

precursor ions were halogen adducts with few structural information. Hence, there is a strong need for an alternative fragmentation method associated to APPI in the negative mode. Irradiation experiments or ion/ion reactions on trapped ions might help to solve this problem [40,41]. 4. Conclusions In summary, successful ionization of various synthetic oligomers (PS, PEG, PMMA, PDMS) was achieved using the APPI technique in the negative ion mode. Experiments using a synchrotron radiation had led us to a better understanding of the ionization mechanisms in the particular case of synthetic polymers. We report for the first time LC-APPI-MS experiments on polymers involving SEC and RPLC separations. However, several drawbacks appeared in the same time: unwanted fragmentations in the positive ion mode; the ionization of the chains of higher masses remained poor. It is worth noting that the understanding of the ionization mechanisms is decisive for the most reliable estimation of the average molecular masses and the correct assignment of the ions. As commercial APPI sources typically use Krypton lamps with a fixed energy (10 eV), the study of the ionization of polymer standards using APPI over a wavelength range was a great opportunity. Acknowledgments We acknowledge SOLEIL for provision of synchrotron radiation facilities (Proposal numbers 20100410 and 20110245) and we would like to thank Alexandre Giuliani and Matthieu Réfrégiers for assistance in using beamline DISCO. References [1] H. Pasch, W. Shrepp, in: H.G. Barth, H. Pash (Eds.), MALDI-TOF Mass Spectrometry of Synthetic Polymers, Springer-Verlag, Berlin, Germany, 2003. [2] G. Montaudo, R.P. Lattimer, in: G. Montaudo, R.P. Lattimer (Eds.), Mass Spectrometry of Polymers, CRC Press, Boca Raton, FL, 2002. [3] G. Montaudo, D. Garozzo, M.S. Montaudo, C. Puglisi, F. Samperi, Macromolecules 28 (1995) 7983–7989. [4] G. Montaudo, M.S. Montaudo, C. Puglisi, F. Samperi, Rapid Commun. Mass Spectrom. 9 (1995) 453–460. [5] C. Wesdemiotis, N. Solak, M.J. Polce, D.E. Dabney, K. Chaicharoen, B.C. Katzenmeyer, Mass Spectrom. Rev. 30 (2011) 523–559. [6] Z. Takats, J.M. Wiseman, B. Gologan, R.G. Cooks, Science 306 (2004) 471–473. [7] A.T. Jackson, J.P. Williams, J.H. Scrivens, Rapid Commun. Mass Spectrom. 20 (2006) 2717–2727. [8] J.P. Williams, G.R. Hilton, K. Thalassinos, A.T. Jackson, J.H. Scrivens, Rapid Commun. Mass Spectrom. 21 (2007) 1693–1704.

[9] N. Bonnaire, A. Dannoux, C. Pernelle, B. Amekraz, C. Moulin, Appl. Spectrosc. 64 (2010) 810–818. [10] M. Friia, V. Legros, J. Tortajada, W. Buchmann, J. Mass Spectrom. 47 (2012) 1023–1033. [11] M. Nefliu, A. Venter, R.G. Cooks, Chem. Commun. (2006) 888–890. [12] R.B. Cody, J.A. Laramee, H.D. Durst, Anal. Chem. 77 (2005) 2297–2302. [13] C.N. McEwen, R.G. McKay, B.S. Larsen, Anal. Chem. 77 (2005) 7826–7831. [14] M. Yamashita, J.B. Fenn, J. Phys. Chem. 88 (1984) 4451–4459. [15] C.M. Whitehouse, R.N. Dreyer, M. Yamashita, J.B. Fenn, Anal. Chem. 57 (1985) 675–679. [16] E.C. Horning, M.G. Horning, D.I. Carroll, I. Dzidic, R.N. Stillwel, Anal. Chem. 45 (1973) 936–943. [17] D.I. Carroll, I. Dzidic, R.N. Stillwell, K.D. Haegele, E.C. Horning, Anal. Chem. 47 (1975) 2369–2373. [18] D.B. Robb, T.R. Covey, A.P. Bruins, Anal. Chem. 72 (2000) 3653–3659. [19] J.A. Syage, M.D. Evans, K.A. Hanold, Am. Lab. 32 (2000) 24. [20] S. Keki, L. Nagy, A. Kuki, M. Zsuga, Macromolcules 41 (2008) 3772–3774. [21] S. Keki, J. Torok, L. Nagy, M. Zsuga, J. Am. Soc. Mass Spectrom. 19 (2008) 656–665. [22] L. Nagy, V. Palfi, M. Narmandakh, A. Kuki, A. Nyiri, B. Ivan, M. Zsuga, S. Keki, J. Am. Soc. Mass Spectrom. 20 (2009) 2342–2351. [23] P. Terrier, B. Desmazieres, J. Tortajada, W. Buchmann, Mass Spectrom. Rev. 30 (2011) 854–874. [24] A. Giuliani, J.L. Giorgetta, J.P. Ricaud, F. Jamme, V. Rouam, F. Wien, O. Laprevote, M. Refregiers, Nucl. Instrum. Methods Phys. Res., Sect. A 279 (2012) 114–117. [25] J. Allegrand, D. Touboul, A. Giuliani, A. Brunelle, O. Laprevote, Int. J. Mass Spectrom. 321 (2012) 14–18. [26] A. Bagag, A. Giuliani, M. Refregiers, F. Le Naour, Int. J. Mass Spectrom. 328 (2012) 23–27. [27] J. Falkenhagen, S. Weidner, Hyphenated techniques, in: C. Barner-kowollik, T. Gruending, J. Falkenhagen, S. Weidner (Eds.), Mass Spectrometry in Polymer Chemistry, Wiley-VCH Verlag & Co, Weinheim, Germany, 2012, pp. 209–235. [28] A. Giuliani, F. Jamme, V. Rouam, F. Wien, J.L. Giorgetta, B. Lagarde, O. Chubar, S. Bac, I. Yao, S. Rey, C. Herbeaux, J.L. Marlats, D. Zerbib, F. Polack, M. Refregiers, J. Synchrotron Radiat. 16 (2009) 835–841. [29] I.V. Chernushevich, B.A. Thomson, Anal. Chem. 76 (2004) 1754–1760. [30] T.J. Kauppila, T. Kuuranne, E.C. Meurer, M.N. Eberlin, T. Kotiaho, R. Kostiainen, Anal. Chem. 74 (2002) 5470–5479. [31] T.J. Kauppila, T. Kotiaho, R. Kostiainen, A.P. Bruins, J. Am. Soc. Mass Spectrom. 15 (2004) 203–211. [32] P.J. Linstrom, W.G. Mallard, NIST Chemistry WebBook, NIST Standard Reference Database Number 69, 2009. [33] E. Basso, E. Marotta, R. Seraglia, M. Tubaro, P. Traldi, J. Mass Spectrom. 38 (2003) 1113–1115. [34] A. Delobel, S. Roy, D. Touboul, K. Gaudin, D.P. Germain, A. Baillet, F. Brion, P. Prognon, P. Chaminade, O. Laprevote, J. Mass Spectrom. 41 (2006) 50–58. [35] S. Roy, A. Delobel, K. Gaudin, D. Touboul, D.P. Germain, A. Baillet, P. Prognon, O. Laprevote, P. Chaminade, J. Chromatogr. A 1117 (2006) 154–162. [36] L.G. Song, A.D. Wellman, H.F. Yao, J.E. Bartmess, J. Am. Soc. Mass Spectrom. 18 (2007) 1789–1798. [37] E. Marotta, R. Seraglia, F. Fabris, P. Traldi, Int. J. Mass Spectrom. 228 (2003) 841–849. [38] K.M. Kadish, J.E. Anderson, Pure Appl Chem. 59 (1987) 703–714. [39] A. Giuliani, D. Debois, O. Laprevote, Eur. J. Mass Spectrom. 12 (2006) 189–197. [40] L. Sleno, D.A. Volmer, J. Mass Spectrom. 39 (2004) 1091–1112. [41] V. Scionti, C. Wesdemiotis, Tandem mass spectrometry analysis of polymer structures and architectures, in: C. Barner-kowollik, T. Gruending, J. Falkenhagen, S. Weidner (Eds.), Mass Spectrometry in Polymer Chemistry, Wiley-VCH Verlag & Co, Weinheim Germany, 2012, pp. 57–84.