Anal Bioanal Chem (2008) 392:595–607 DOI 10.1007/s00216-008-1969-0

ORIGINAL PAPER

Characterization of linear and branched polyacrylates by tandem mass spectrometry Kittisak Chaicharoen & Michael J. Polce & Anirudha Singh & Coleen Pugh & Chrys Wesdemiotis

Received: 7 January 2008 / Revised: 7 February 2008 / Accepted: 9 February 2008 / Published online: 30 March 2008 # Springer-Verlag 2008

Abstract The unimolecular degradation of alkali-metal cationized polyacrylates with the repeat unit CH2CH (COOR) and a variety of ester pendants has been examined by tandem mass spectrometry. The fragmentation patterns resulting from collisionally activated dissociation depend sensitively on the size of the ester alkyl substituent (R). With small alkyl groups, as in poly(methyl acrylate), lithiated or sodiated oligomers (M) decompose via freeradical chemistry, initiated by random homolytic C-C bond cleavages along the polymer chain. The radical ions formed this way dissociate further by backbiting rearrangements and β scissions to yield a distribution of terminal fragments with one of the original end groups and internal fragments with 2–3 repeat units. If the ester alkyl group bears three or more carbon atoms, cleavages within the ester moieties become the predominant decomposition channel. This distinct reactivity is observed if R=t-butyl, n-butyl, or the mesogenic group (CH2)11-O-C6H4-C6H4-CN. The [M+alkali metal]+ ions of the latter polyacrylates dissociate largely by charge-remote 1,5-H rearrangements that convert COOR to COOH groups by expulsion of 1-alkenes. The acid groups may displace an alcohol unit from a neighboring ester pendant to form a cyclic anhydride, unless hindered by steric effects. Using atom transfer radical polymerization, hyperbranched polyacrylates were prepared carrying ester groups

K. Chaicharoen : M. J. Polce : C. Wesdemiotis (*) Department of Chemistry, The University of Akron, Akron, OH 44325-3601, USA e-mail: [email protected] A. Singh : C. Pugh Department of Polymer Science, The University of Akron, Akron, OH 44325-3909, USA

both within and between the branches. Unique alkenes and alcohols are cleaved from ester groups at the branching points, enabling determination of the branching architecture. Keywords Polyacrylate fragmentation . Degradation mechanisms . Tandem mass spectrometry . Mesogenic substituents . Hyperbranched polyacrylates . Branching architecture . Charge-remote rearrangement

Introduction Poly(alkyl acrylate)s are readily ionized by matrix-assisted laser desorption ionization (MALDI [1, 2]) and electrospray ionization (ESI [3]) [4]. Both are soft ionization methods, generating intact molecular ions, usually of the type [M+X]+, X=alkali metal, from each oligomer (M) [4, 5]. MALDI yields singly metalated ions whereas ESI may lead to a distribution of charge states depending on the size of the polymer; singly metalated ions predominate for oligomers with <∼15 repeat units [4, 6, 7]. MALDI and ESI mass spectra reveal the mass-to-charge ratios and relative abundances of the oligomers present in the polymer sample. For linear polyacrylates, these data provide information about the distribution of chain lengths and combined end groups produced in the polymerization [4–14]. For further insight into the precise connectivity of individual polymer chains, tandem mass spectrometry (MS-MS) experiments are needed [7, 15–26]. Here, a single oligomer ion is selected and induced to form fragments, most frequently by collisionally activated dissociation (CAD). The resulting fragmentation patterns reveal individual chain-end [7, 15–25], in-chain [25], or side-chain substituents [26]. MS-MS studies have been reported for several poly(alkyl methacrylate)s [7, 15– 22, 27]. Here, we report first MS-MS data for poly(alkyl

596

Anal Bioanal Chem (2008) 392:595–607

acrylate)s with various alkyl side chains and architectures and examine in detail their fragmentation pathways. Linear, side-chain liquid crystalline polyacrylates can be prepared by atom-transfer radical polymerization (ATRP) methods [28] using monomers with mesogenic ester groups [29–32]. If these monomers also contain an initiating site, i.e. if they are “inimers”, hyperbranched architectures with liquid-crystal properties are generated [33, 34]. As will be shown in this study, linear and branched architectures can be distinguished based on the fragmentations of polyacrylate [M+alkali metal]+ ions.

methyl, 1,400 for the t-butyl, and 1,400 for the n-butyl polyacrylate, and the polydispersity indices (Mw/Mn) were 1.13 for all three standards. All purchased chemicals were used in the condition received. The side-chain liquid crystalline polymers were produced by atom-transfer radical polymerization by Pugh and coworkers, as summarized in Fig. 1. The linear polyacrylate was obtained by reaction of a brominated initiator with the monomer according to the pathway shown in Fig. 1a [29]. It contains the repeat unit C27H33O3N (419.25 Da) in the connectivity H-[C27H33O3N]n-Br. The hyperbranched polymer was synthesized by using an acrylate monomer that also contained a chloroalkyl group which can act as an initiating and/or propagating site (Fig. 1b) [33–35]; such monomers are termed inimers because they can self-initiate a polymerization. Starting with the dimer, each inimer (or any oligomer) can be added at more than one site, leading to differently branched oligomers with the repeat unit C30H36O5NCl (525.23 Da) and the overall composition [C30H36O5NCl]n. Gel permeation chromatography (GPC) with light-scattering detection indicated that the polyacrylates with mesogenic side chains had Mn and Mw/Mn values of 4700 and 2.49, respectively. The polymers were precipitated in cold methanol and, subsequently, dissolved in THF and reprecipitated in cold methanol several times before GPC and mass spectral characterization.

Experimental Materials The matrix, salts, and solvents for the mass spectrometric analyses were purchased from Aldrich (Milwaukee, WI, USA) or VWR (West Chester, PA, USA). Poly(methyl acrylate), poly(t-butyl acrylate), and poly(n-butyl acrylate) standards with the connectivity C4H9-[CH2CH(COOR)]n-H (R=CH3, t-C4H9, and n-C4H9, respectively) were purchased from Polymer Source (Dorval, Quebec, Canada). The average molecular weights (Mn) were 2,800 for the

CH3

a

CH C

O

+

C

OR

CH2

C O

Cu(I) catalyst

O

ATRP

monomer

CH2

CH

CH2

CH

Br n-1

C O

C O

OR

OR

CH C

O

O

CH2

CH

CH2

COOR

CH C

Cl O

O

COOR

propagation sites

inimer

CH

linear polymer

CH Cl ATRP

CH2

CH3

O

OR

initiator

b

Cu(I) catalyst

n CH2 CH

Br

CH2

CH Cl COOR

hyperbranched polymer dimer

c

R=

(CH2)11 O

CN

mesogenic group in ester side chain

Fig. 1 (a) Linear [29–31] and (b) hyperbranched [33, 34] polyacrylates synthesized by ATRP; (c) mesogenic ester group included in these polyacrylates

Anal Bioanal Chem (2008) 392:595–607

MALDI mass spectrometry MALDI-MS-MS experiments were performed on a Waters Q/ToF Ultima quadrupole/orthogonal-acceleration timeof-flight mass spectrometer (Milford, MA, USA), equipped with a pulsed nitrogen laser emitting at 337 nm. Solutions of dithranol matrix (20 mg mL−1), polymer (10 mg mL−1), and lithium trifluoroacetate cationizing agent (10 mg mL−1) were mixed in the ratio 10:2:1, and approximately 1.0 μL of the final mixture was deposited on the 96-well sample holder plate that is inserted into the MALDI source. The [M+Li]+ ions exiting the MALDI source were directed toward the quadrupole mass filter, which was set to transmit one oligomer mass only (mass-selective mode). The precursor ion resolution was adjusted to select the complete isotopic cluster. The selected ion proceeded to an rf-only hexapole collision cell, pressurized with Ar at ∼0.9–1.0 bar, where CAD took place (multiple collision conditions). Laboratory-frame kinetic energies were adjusted within 80–110 eV to give center-of-mass collision energies of ∼2 eV for all precursor ions. The fragments and undissociated precursor ions exiting the collision cell were focused through an rf-only hexapole lens and the focused ion packet was accelerated orthogonally by ∼10 kV into the ToF region for mass analysis. Control mass spectra were measured with the ToF mass analyzer by setting the quadrupole mass filter to rf-only mode, so that it transmitted all ions produced in the MALDI source. MALDI mass spectra were also acquired on a Bruker Reflex III ToF mass spectrometer (Billerica, MA, USA) following an analogous procedure, which has been described in detail elsewhere [36]. The latter spectra showed much less in-source fragmentation and, thus, were more suitable for the analysis of the sample purity. Cationization by Na+ gave rise to very similar mass spectra; the [M+Li]+ adducts fragmented more efficiently, however, presumably because of the higher binding energy of Li+ vs. Na+. The ion abundances of several ToF scans were summed to obtain mass or tandem mass spectra with good signal-to-noise ratio. Quoted m/z values are monoisotopic. ESI mass spectrometry ESI-MS-MS experiments were performed on a Bruker Esquire-LC quadrupole ion-trap mass spectrometer (Billerica, MA, USA) equipped with electrospray ionization. THFmethanol (1:1) solutions of the polyacrylates were doped with 1 μL sodium iodide solution in THF (10 mg mL−1). [M+Na]+ ions were formed by spraying the mixed solution into the ion source with a syringe pump at a rate of 350 μL h−1. The spraying needle was grounded and the entrance of the sampling capillary was set at −4 kV for the analysis of positive ions. Nitrogen was used as the nebulizing gas (10 psi) and drying gas (8 L min−1, 300 °C), and He was the

597

bath gas in the ion trap (10−8 bar). The desired polyacrylate precursor ion ([M+Na]+) was isolated and induced to fragment by CAD inside the trap by resonance excitation with a radiofrequency (rf) field. The excitation time was set at 40 ms and the rf amplitude (Vp-p) was adjusted in the 1.20–2.50 V range to maximize fragment ion abundances. MS3 spectra were acquired similarly by isolating one MS-MS fragment in the trap and inducing its fragmentation by subsequent CAD; fifty to sixty-five scans per spectrum were averaged. Quoted m/z values are monoisotopic.

Results and discussion Poly(methyl acrylate), PMA Figure 2 shows the MALDI-CAD spectrum of the lithiated 22-mer from the PMA standard C4H9-[CH2CH(COOCH3)]n-H (86-Da repeat unit). For simplicity, the COOCH3 ester group will be abbreviated by E. The spectrum contains two major distributions of fragment ions, viz. an and yn, corresponding to the Li+ complexes of truncated PMA chains with the composition C4H9-[CH2CH(E)]n−2-CH2C(E)=CH2 (an) or CH2 =C(E)-[CH2CH(E)]n−1-H (yn). These fragments carry one of the original end groups (α or ω) and a new =CH2 end group, and appear at m/z (n−2)86+163 or m/z (n−1)86+93, respectively. Superimposed on the an/yn distributions are the abundant low-mass fragments at m/z 166 and 265, which are assigned to the internal radical ion [•CH(E)CH2CH(E)-H+Li]+ and the internal closed-shell ion [CH2 = C(E)-[CH2CH(E)]2-H+Li]+, respectively. The CAD fragmentation pattern of lithiated PMA is similar to that reported recently for silverated polystyrene [23, 24]. Since both polymers have polyethylene frames with short side-chain substituents, it is not surprising that their metal-cationized forms dissociate through analogous channels when energetically excited. Silverated polystyrene oligomers decompose via free radical chemistry pathways, commencing with random homolytic C-C bond cleavages along the polymer backbone [23, 24]. The same chemistry can be invoked to rationalize the fragments generated upon CAD of [PMA+Li]+ oligomers, as outlined in Figs. 3 and 4. Random C-C bond homolyses in the chains of [PMA+Li]+ oligomers generate
charged radicals containing
either the α end group a•n ; b•n or the ω end group y•n ;
z•n , cf. Fig. 3b. Two of these radicals are primary a•n ; y•n
, and two are secondary resonance-stabilized radicals b•n ; z•n . This nomenclature is briefly outlined in Fig. 3a; it was recently introduced to designate MS-MS fragments from ionized polymer chains in a meaningful manner [23], similar to the one used for fragments from protonated peptides [37, 38]. The actually observed fragments arise by consecutive dissociations of the

598

Anal Bioanal Chem (2008) 392:595–607

Fig. 2 MALDI-CAD mass spectrum of the [M+Li]+ ion of the poly(methyl acrylate) 22-mer (repeat unit, 86.04 Da)

[M + Li]+ 265

1958

K3

895

y11b

879

867

y10 a11b a10

.

166

851

J2

y4 y2 a3

y6 a5

y8

a7

200

y12

a9

a14

1000

600

y3

a

H2 C

C 4H 9

y2

z3

z2

y1

H2 C

H C

b1

1800

m/z

H C

H

E

b2

a2

y20

ω

z1

E

a1

a18

140 1400

H2 C

H C

E

α

y16

a3

b3

b Li+

an

•

C 4H 9

CH2

CH

bn•

CH2 CH CH2 n-2

E

Li

CH

E

CH2

E

CH2 CH

CH2 CH

zn•

H

CH

E

n-1

Li+

C 4H 9

CH2 CH

n-1

E

Li

CH2 CH E

+

H

yn•

n-1

E

E

+

c Li+

an

C 4H 9

CH2

CH2 C

CH E

n-2

Li+

CH2

CH2

C E

E

H

CH2 CH

n-1

E

+

+

Li

anb

C4H9

CH2

CH E

bn

CH2 CH

Li

CH2

CH2

CH

n-2

CH2 CH

H

CH2 CH

E

E

CH n-1

CH E

Fig. 3 (a) Nomenclature for backbone fragments from poly(methyl acrylate) ions [23]. E abbreviates the COOCH3 group. (b) The four types of radical ion formed by random homolytic C-C bond cleavages

ynb

n-1 Li+

Li+

C4H9

yn

CH E

CH

CH E

CH2

CH E

H

zn

n-2

in the poly(methyl acrylate) chain. (c) Closed-shell fragments produced by consecutive dissociations of the radical ions shown in (b)

Anal Bioanal Chem (2008) 392:595–607 Fig. 4 Backbiting via 1,5-H rearrangement and consecutive β C-C bond scissions in the secondary radical ions formed by random homolytic C-C bond cleavages in the poly(methyl acrylate) chain

599 E HC

Li

CH 2

H R

CH 2

CH E

CH 2

C

n

+

E

C H2

R

CH 2

CH E

bn• / zn•

Li+

CH E

Li

CH

CH 2

E

CH 2 n

CH 2 n

CH 2

E

E

C

CH 2

E

C

CH 2

Li

CH

incipient radicals, which involve typical radical-induced reactions, viz. bond cleavages in the position β to the unpaired electron and 1,5-H (backbiting)rearrangements [39]. The primary radical ions a•n y•n can undergo β-H• loss to yield the closed-shell fragments an/yn (Fig. 3c) or one of two possible β C-C bond scissions, releasing either the • E radical (i.e. •COOCH3) to form the closed-shell fragments  anb/ynb (Fig. 3c) or a monomer unit to form a•n 1 y•n 1 (depolymerization). Using the primary radical • CH2CH(COOCH3)CH3 as a model for a•n y•n, the activation energies for β scissions of H•, •COOCH3, and methyl acrylate monomer are estimated at ca. 130, 140, and 70 kJ mol−1, respectively [40–43]. Clearly, depolymerization is the energetically  most favorable process. Complete depolymerization of a•n y•n to [monomer+Li]+ (m/z 93) is feasible under the multiple collision conditions of our MALDI-CAD experiments; however, this fragment is not observed, because it falls under the low-mass cutoff of the quadrupole mass filter.  Dissociation of a•n y•n via β loss of H• or •COOCH3 is associated with approximately twice as much critical energy as dissociation via monomer evaporation. Consistent with their high energy requirement, series anb/ynb (elimination of • COOCH3) have very low relative abundances. Since the critical energies for β-H• and β-•COOCH3 losses from a•n y•n are very similar, the fraction of series an/yn generated from a•n y•n (elimination of H•) must also be small, strongly suggesting that there is an alternative route to the dominant an/yn series.

E

K3 (m/z 265)

+

+

R

CH 2

CH 2

CH E

E

an / yn

+

CH 2

CH 2

E

internal fragments Li

E

CH

CH 2

E

CH

CH 2

+

+ CH 2

E

CH 2

J2• (m/z 166) R

C

smaller

bn• / zn•

CH 2 n-1

CH E

(depolymerization)

The direct bond cleavages available to the secondary • • evaporation to yield smaller radical  •ions bn zn are monomer • bn 1 zn 1 radicals and β-H• loss to yield the closed-shell fragments bn/zn (Fig. 3c). The corresponding activation energies are estimated at ca. 120 and 160 kJ mol−1, respectively, if the secondary radical •CH(COOCH3)CH2CH3 is used as a model for bn• z•n [40–43]. Fragments bn/zn are at or below the noise level, in agreement with their considerable critical energies. An energetically far more favorable reaction in this case is a 1,5-H rearrangement via a six-membered ring transition state (backbiting), followed by β C-C bond scissions within the polymer chain, as shown in Fig. 4. Based on available literature data, the first step costs approximately 30 kJ mol−1 [44, 45], and the overall reaction sequence −1 approximately 70 kJ mol  • [40–45]. Backbiting converts the • incipient secondary bn zn radical ions to more stable tertiary species that can further decompose by two competitive β C-C bond scissions (Fig. 4). One coproduces the terminal series an/yn and the internal radical on J•2 (m/z 166), and the other leads to the internal trimer K3 (m/z 265) or a truncated secondary radical ion; the latter can undergo backbiting anew and consecutive β C-C bond cleavages,  or dissociate via monomer evaporations. Note that all b•n z•n radical ions give rise to the same internal J•2 and K3 fragments, reconciling the large relative intensity of these ions. The backbiting pathway of Fig. 4 is assumed to be the major route to the abundant terminal fragments an/yn, as it requires a significantly lower  barrier than the direct β-H• scission from a•n y•n (vide supra).

Anal Bioanal Chem (2008) 392:595–607

1250

1418

1306

Fig. 5 MALDI-CAD mass spectrum of the [M+Li]+ ion of the poly(t-butyl acrylate) 11-mer (repeat unit, 128.08 Da)

1362

600

1138

1474

1194

[M + Li]+

- 56 - 56 - 56 - 56

- 56 - 56 - 56 - 56 - 56 - 56 - 56 600

800

1000

Poly(t-butyl acrylate), PtBA Completely different fragmentation behavior is encountered with [M+Li]+ ions from PtBA oligomers, which are found to dissociate exclusively at their side chains. This is attested in Fig. 5 by the CAD spectrum of the lithiated 11-mer. The only significant fragmentation channel taking place is sequential eliminations of isobutene units (56 Da) from the t-butyl acrylate substituents. These reactions probably proceed by hydrogen transfer from the ester alkyl group to the carbonyl oxygen and concerted β elimination of isobutene via an energetically preferred six-membered ring, as shown in Fig. 6a; they can be classified as sigmatropic, charge-remote rearrangements. Under the multiple collision conditions of our CAD experiments, many sequential isobutene eliminations occur; for example, the 11-mer loses up to eleven C4H8 units, cf. Fig. 5.

a C4H9

1200

It is noteworthy that no backbone cleavages take place at the same time, indicating that the consecutive charge-remote C4H8 losses are associated with substantially lower barriers than competitive dissociations via homolytic bond cleavages in the polyacrylate chain. Poly(n-butyl acrylate), PnBA The fragmentation patterns of [PnBA+Li]+ and [PtBA+Li]+ ions are fairly similar but not identical. Both primarily dissociate by reactions occurring in their side chains. Essentially one process, viz. C4H8 losses from the t-butyl ester pendants, is observed from [PtBA+Li]+ oligomers (Fig. 5). In contrast, [PnBA+Li]+ oligomers undergo two competitive fragmentations, involving the elimination of either 1-butene (56 Da) or 1-butanol (74 Da) from the n-butyl ester pendants (Fig. 7). The former reaction can be rationalized by a

Li+

Li+ CH2 CH

n-1

CH2 CH H

C O

C

OC(CH3)3

O

O C H3C

C4H9

CH2 CH

H CH2

n-1

CH2 CH H

C O

C O

OC(CH3)3

OH

CH2

+

C

56 Da

Li+

Li+ CH2

CH2 CH2

CH

n-2

CH2

C O O(CH2)3CH3

CH

O C

CH3

H3C

CH3

b C4H9

m/z

1400

O O H (CH2)3CH3

CH H C O

C4H9

CH2

CH

n-2

CH2

C O O(CH2)3CH3

CH H

CH

C

C O +

O

O

HO(CH2)3CH3

74 Da Fig. 6 (a) Charge-remote 1,5-H rearrangement in the t-butyl ester side chain of poly(t-butyl acrylate). The reaction is illustrated at the terminal repeat unit, but can take place at any ester pendant. (b)

Intramolecular nucleophilic displacement of an alcohol molecule from the ester group adjacent to the COOH group created in (a)

601 1474

Anal Bioanal Chem (2008) 392:595–607

- 56

1418

1-butene losses may follow. Alternatively, the newly created acid can attack an adjacent ester group to release 1-butanol (74 Da) and form a cyclic anhydride (Fig. 6b). Consistent with this sequence of events, 1-butanol elimination proceeds with appreciable yield only after prior 1-butene loss, cf. Fig. 7. The intramolecular anhydride formation is not observed with [PtBA+Li]+ oligomers. The absence of this channel with t-butyl acrylate ester pendants is attributed to the steric hindrance imposed by the bulky t-butyl groups, which could prevent an adjacent acid from attacking its neighboring t-butyl ester groups to displace t-butanol.

- 56

1362

- 74

- 56

- 56

1000 1200 1400 m/z Fig. 7 MALDI-CAD mass spectrum of the [M+Li]+ ion of the poly (n-butyl acrylate) 11-mer (repeat unit, 128.08 Da)

Linear polyacrylate with mesogenic ester groups

charge-remote H rearrangement analogous to the one discussed for poly(t-butyl acrylate), cf. Fig. 6a. The latter reaction is attributed to an intramolecular displacement leading to an anhydride, as depicted in the scheme of Fig. 6b. Initially, 1-butene (56 Da) is eliminated, creating a free acid group at the side chain (Fig. 6a). Consecutive

The synthetic method used for this polyacrylate gives rise to a secondary bromine substituent at the ω chain end (Fig. 1a). This polymer is not observed in the MALDI mass spectrum (Fig. 8a); the dominant distribution in the latter spectrum, A, arises from the corresponding dehydrobromination product which has the connectivity CH3CH(COOR)-

B

A

A

A

B

A

CH

CH

C O

C O

OR

OR

OR

A

B

A

B

B

4000

A B m/z

2124.0 2125.0

+

n

3174

2649

2124

C30H36O5NCl

2126.0 2127.0

c Na

2000

CH

C O

3000

b

1599

CH2

n-2

B

2000

1000

Na+ CH

CH3

2128.0

1281 1353

B

A

B

2538 2610

1700 1772

A

2119 2191

A

a

4000

Fig. 8 MALDI-ToF mass spectra of (a) the linear polyacrylate H[C 27 H 33 O 3 N] n -Br and (b, c) the hyperbranched polyacrylate [C30H36O5NCl]n synthesized by ATRP as shown in Fig. 1a and b,

6000

2129.0 2130.0 2130.9 2131.9 2132.9

- 56

1214 1232

1158

- 74

1288

1344

[M + L i]+

m/z

respectively. (c) Expanded trace of the m/z 2123–2134 region, showing the measured isotopic cluster of the 4-mer from the hyperbranched polymer

Anal Bioanal Chem (2008) 392:595–607

1043.8

Fig. 9 MALDI-CAD mass spectrum of the [M+Li]+ ion of CH3CH(COOR)-[CH2CH (COOR)]3-CH=CH(COOR) (5-mer; repeat unit, 419.25 Da). The structure of R is shown in Fig. 1c

1391.1

602

[M + Li]+

-365 1363.1

-347

500 [CH2CH(COOR)]n−2-CH=CH(COOR). R abbreviates the mesogenic functionality shown in Fig. 1c. Because α and ω end groups add up to two monomer units, the composition of the distribution A oligomers is [CH2CH(COOR)]n (no nominal end groups). A second, less abundant distribution, B, appears 347 Da below A and is due to a dehydrobrominated polyacrylate, in which one COOR ester group has been hydrolyzed to the COOH free acid. Two further unidentified, minor distributions are present in the spectrum and presumably originate from synthetic byproducts, not fragmentation upon MALDI, as they are absent in the CAD spectrum (discussed below). The CAD spectrum of the lithiated 5-mer from distribution A is illustrated in Fig. 9. The fragmentations observed resemble those taking place from lithiated PnBA (cf. Fig. 7). In both cases, the major dissociation products result from eliminations of the alkyl moiety of the ester pendants in the

1000

2104.7

1409.1

1015.8

-347

1756.4

-347 845.7

-347

678.5

331.2

426.4

696.5

-365

1500

-347

2000 m/z 200

form of either an alkene or alcohol (Fig. 6). Table 1 lists the compositions and masses of the alkene and alcohol losses from the polyacrylate with mesogenic ester groups. As with PnBA, alcohol elimination (−365 Da) takes place appreciably after a free acid (i.e. COOH) group has been created by prior alkene elimination (−347 Da). The similarity between the fragmentation characteristics of PnBA (Fig. 7) and the mesogenic linear polyacrylate (Fig. 9) is not surprising, considering that both polymers contain primary alkyl ester groups and identical chain structures. A further noticeable common characteristic of these polyacrylates is that they do not undergo 1,5-H rearrangements involving H atoms of the backbone; such rearrangements occur exclusively within the ester functionality. In addition to the major fragments mentioned, several minor dissociation products are detected, especially at low masses, and attributed to consecutive fragmentations of the

Table 1 Alkenesa and alcoholsb eliminated from collisionally activated [polyacrylate+Li]+ ions carrying the mesogenic ester group R (348.2 Da) shown in Fig. 1c Loss

Observed from

[R-H]a C24H29ON, 347.22 Da

Linear polymer and linear isomers or linear segments of hyperbranched polymerc Linear polymer and linear isomers or linear segments of hyperbranched polymerc Linear isomers or linear segments of hyperbranched polymerc Linear isomers or linear segments of hyperbranched polymerc Branched isomers (e.g., B2, B3, B5) of hyperbranched polymer Branched isomers (e.g., B2, B3, B5) of hyperbranched polymer Branched isomers (e.g., B4, B5) of hyperbranched polymer

R-OHb C24H31O2N, 365.24 Da CH2 =CH-COORa C27H33O3N, 419.25 Da HO-CH2CH2-COORb C27H35O4N, 437.26 Da CH2 =C(COOR)-CH2CH2(X)a,d C57H70O8N2, 910.51 Da HO-CH2CH(COOR)-CH2CH2(X)b,d C57H72O9N2, 928.52 Da CH2 =C(COOR)-CH2CH(X)-CH2CH2(X)a,d C87H107O13N3, 1401.78 Da CH2 =CH-COOHa C3H4O2, 72.02 Da a

Hyperbranched polymer (at initiating unit)

Alkenes eliminated according to the pathway shown in Fig. 6a Alcohols eliminated according to the pathway shown in Fig. 6b c Cleaved from the ester side chains of the linear polyacrylate (Fig. 1a) or from the ester side chains of the linear segments (branches) of the hyperbranched polyacrylate (Fig. 1b) d X=COO-CH2CH2-COOR b

Anal Bioanal Chem (2008) 392:595–607

603

isotopic patterns observed for the different n-mers provide evidence that all expected Cl atoms are included in the detected ions. For example, the isotopic cluster and isotope mass-to-charge ratios measured for the 4-mer, Fig. 8c, match, within experimental error, those calculated for four repeat units, confirming the presence of four Cl atoms. This result undoubtedly indicates that secondary Cl atoms introduced by ATRP can be detected in MALDI experiments. In contrast, secondary Br atoms were undetectable (vide supra). A second, minor distribution in the MALDI mass spectrum of the hyperbranched polyacrylate is observed 72 Da below the main distribution and corresponds to polymer that has lost an acrylic acid molecule from the initiating unit (cf. dimer structure in Fig. 1b). Starting with the trimer, each n-mer of the main product contains more than one isomer, each having a unique branching architecture. The isomers possible with the 4-mer are depicted in Fig. 10. The first two monomers can combine in only one fashion, the third monomer can be

major fragments. Two relatively abundant fragments that do not appear to originate from the pathways outlined in Fig. 6 are the ions at m/z 426.4 and 845.7, which are the lithiated monomer and dimer, respectively; a weak lithiated trimer is also observed at m/z 1265.0. It is difficult to discern how these fragments are formed; they could arise by homolytic C-C bond scissions along the chain, giving rise to incipient radicals held together by the metal ion [26], which recombine after monomer evaporation. Hyperbranched polyacrylate with mesogenic end groups ATRP of the inimer employed in the synthesis of the hyperbranched polyacrylate yields a polymer with the composition [C30H36O5NCl]n (cf. Fig. 1b). The formation of such a product is affirmed by the MALDI mass spectrum, which shows a dominant distribution of [C30H36O5NCl]n oligomers with the expected repeat unit of 525 Da, cf. Fig. 8b. Because each monomer contributes one Cl atom, an n-mer should contain n Cl atoms. The

H2C CH C O

Cl

O CH2 CH

CH2 CH

CH2 CH

C O

COOR

O CH2

CH2 CH

C O

Cl CH

O CH2

COOR

Cl

C O

Cl CH

O CH2

COOR

Cl CH COOR

Cl

Cl

Cl

L1

Cl CH2 CH

H2C CH C O O CH2 CH

CH2 CH

C O Cl O CH2 CH

Cl

COOR

C O

COOR

Cl

Cl

Cl CH2 CH

O CH2 CH

C O

COOR

O CH2

Cl CH

B2

Cl

COOR

Cl

Cl

Cl

Cl Cl

Cl

B3

Cl

Cl

Cl

B4

Fig. 10 Linear (L1) and branched (B2-B5) isomers of the 4-mer of the hyperbranched polyacrylate synthesized by ATRP as shown in Fig. 1b. Complete structures are given only for L1 and B2. The first

Cl

Cl

B5

Cl

two repeat units are drawn in black, the third repeat unit in red, and the fourth in blue. The bonds between the repeat units are shown as elongated dashed lines

604

Anal Bioanal Chem (2008) 392:595–607

the sites marked * in Fig. 11, cleaves only pieces of 347 or 419 Da from the ester pendants (cf. Table 1). If this rearrangement takes place at the ester linkage of a branch, for example the encased site of B2 in Fig. 11, a much larger piece is released, and its mass (911 Da from B2) identifies the branch length and possible branch position. Also the alcohol displacements cleave larger units if they take place at branch positions, cf. Table 1. The CAD spectrum of the lithiated 4-mer from the reduced hyperbranched polyacrylate, Fig. 12, contains abundant fragments from eliminations of 419-Da and 437-Da molecules. These reactions can only proceed from unbranched monomer units and, hence, must originate from the linear isomer L1 or from linear segments of the branched isomers, such as the two terminal repeat units of B2. The consecutive losses of an alkene (419 Da), alcohol (437 Da), and alkene (419 Da) from three ester side chains are best reconciled with the linear architecture L1; an analogous reaction sequence dominated during CAD of the linear mesogenic polyacrylate, cf. Figs. 12 and 9. There are also fragments characteristic of branched structures in the CAD spectrum of Fig. 12. For example, the sizable fragment at m/z 1061.4 corresponds to the elimination of a 911-Da alkene from [M+Li]+ which, as mentioned above, is accounted for by a 1,5-H rearrangement that cleaves a branch. The ion at m/z 1061.4 can arise from architectures B2 (Fig. 11), B3, and B5. The alkene cleaved is observed as lithiated ion at m/z 917.2. On the other hand, architectures B4 and B5 are likely sources of the fragment at m/z 570.2, which corresponds to the elimination of a 1402Da alkene (Table 1); the latter is observed in ionized form at m/z 1409.5. Note that the alkenes generated in the 1,5-H rearrangements give consistently less intense peaks than the

attached at two different sites, and the fourth at three different sites. Depending on where the third and fourth monomers are added, five distinct isomers can emerge, one linear (L1 in Fig. 10) and four branched (B2-B5). The term linear describes the architecture in which all monomer units (except the initiating unit) carry COO-CH2CH(C1)-COOR side chains; all other architectures are termed branched (cf. [46] for a more descriptive notation for branched polyacrylate structures). Neither NMR nor MS-MS experiments could establish with certainty that branched architectures were prepared. CAD of lithiated [C30H36O5NCl]n causes multiple HCl losses from [M+Li]+, as well as 347-Da alkene losses from the mesogenic ester groups and acrylic acid evaporations (both coupled with sequential HCl losses). These reactions corroborate the composition of the polymer but do not provide connectivity insight. For more informative MS-MS analysis, all Cl were reduced to H atoms. Isomers L1 and B2 of the 4-mer from the reduced polymer are shown in Fig. 11. The new repeat unit, C30H37O5N (491.27 Da), is closely related to the monomer of the linear mesogenic polyacrylate discussed in the forgoing section, C27H33O3N (419.25 Da), differing from it by one extra acrylic acid molecule in the esterified form-compare the structure in Fig. 8a with those in Fig. 11. Polyacrylates with longer alkyl groups in their ester pendants, like those included in Fig. 11, mainly fragment by charge-remote 1,5-H rearrangement and alcohol displacement at the ester moiety (vide supra and Figs. 7 and 9). Such reactions can take place at two different locations within each reduced repeat unit of the hyperbranched polymer. A 1,5-H rearrangement taking place in the linear isomer or within linear segments of a branched isomer, for example at Fig. 11 Reduced forms of isomers L1 and B2 of the hyperbranched polyacrylate synthesized according to Fig. 1b. The asterisks indicate ester groups in linearly connected monomer units. The ester moiety at the branched unit in B2 is encased

H2C CH C O O CH2 CH

CH2 CH

CH2 CH

COOR

C O

*

O CH2 CH2

L1

C O

C O

O CH2 CH2

O CH2 CH2

COOR

*

COOR

COOR

CH2 CH2

H2C CH

*

C O O CH2 CH COOR

B2

CH2 CH2

CH2 CH

C O O CH2 CH2

*

C O O CH2 CH COOR

COOR CH2 CH2 C O O CH2 CH2 COOR

Anal Bioanal Chem (2008) 392:595–607

605

[M + L i]+

800

1971.7 1624.6

-34 7

1409.5 ∅

1899.7

1480.6

989.4

-34 7 1277.4

858. 2

917.2 ∅

768. 2

-36 5

-419

-41 9 1205.4

1061.4 ∅

624.2

498.2

570.2 ∅

349.0 378.2

277.0

444.2

-41 9

400

-43 7

1115.4 1133.4

696. 2

1552.6

Fig. 12 MALDI-CAD mass spectrum of the [M+Li]+ ion of the 4-mer from the reduced hyperbranched polymer, [C30H37O5N]n (repeat unit, 491.27 Da). A superscripted ∅ next to the m/z value marks peaks assigned to branched isomers

120 0

160 0

m/z

H2C CH C O O CH2 CH

CH2 CH

from B5

C O

COOR

O CH2 CH

H2C CH C O

OH

CH2 CH

CH2 CH

C O

COOR

C O

COOR

from B3

O CH2 CH

CH2 CH2

C O

O CH2 CH2 OH

H2C CH C O O CH2 CH

CH2 CH

COOR

from B2

1077

COOR

[M + Na]+

CH2 CH

C O

C O

OH

O CH2 CH2 COOR

-419

500 3

1059

-347

700

1005

712

-72

658

586

730

-365

1100

m/z

Fig. 13 ESI-CAD (MS ) mass spectrum of the m/z 1077 fragment in the ESI-CAD mass spectrum of the sodiated 4-mer of [C30H37O5N]n (repeat unit, 491.27 Da)

606

complementary free acids (reaction in Fig. 6a), indicating that the carboxylic acid functionality leads to a more strongly bound Li+, perhaps because of the formation of a salt bridge. Very similar trends are observed in the CAD spectrum of the lithiated 5-mer, which also is consistent with a mixture of linear and branched architectures. Thus, tandem mass spectrometry provides convincing evidence that branched mesogenic oligomers have been synthesized. Because of the complexity of the product, only qualitative analysis is feasible in MS-MS mode. A more quantitative assessment of the branched isomer mixture should be achievable by multistage mass spectrometry (MSn). To test this supposition, the structure of the fragment at m/z 1061.4 (loss of 911 Da) was probed by MS3 using ESI in a quadrupole ion trap. The mesogenic polyacrylates ionized less efficiently by ESI than MALDI, and the best signal-to-noise ratio was obtained with Na+ cationization. The m/z 1061 fragment shifts to m/z 1077 when Li+ is replaced by Na+. ESI-CAD of [M+Na]+ (spectrum not shown) and MALDI-CAD of [M+Li]+ give rise to similar fragmentation patterns; the former activation method causes fewer consecutive reactions, but in both cases the loss of 911 Da yields a sizable fragment. The ESI-MS3 spectrum of this fragment, m/z 1077, is reproduced in Fig. 13. The thirdgeneration products present in this spectrum are best rationalized with the structure produced from a 4-mer with the B2 type of branching, which is inset in Fig. 13, together with the m/z 1077 isomers emerging from architectures B3 and B5. The m/z 1077 fragment from B2 can lose 347-Da and 419-Da alkenes, as is indeed observed. Moreover, its COOH group is generated at a location from which it can displace the alcohol R-OH (365 Da); also a 72-Da unit can be cleaved from the initiating end. The elimination of 365 Da would be less favorable in the m/z 1077 product from B3, as it would lead to a strained ten-membered ring anhydride. Conversely, elimination of 419 Da is obstructed in the m/z 1077 product from B5, in which all repeat units are branched. Hence, the MS3 data are most consistent with B2 being the major branched architecture of the 4-mer.

Conclusions Alkali-metal cationized polyacrylates that have been energetically activated decompose by charge-remote reactions. The size of their ester side chains determines whether homolytic bond cleavages in the chain, or sigmatropic rearrangements and nucleophilic displacements within the ester substituents predominate. The former pathways are observed for poly(methyl acrylate) and the latter for poly (t-butyl acrylate), poly(n-butyl acrylate), and side-chain liquid-crystalline polyacrylates with long alkyl chains in their ester groups.

Anal Bioanal Chem (2008) 392:595–607

An important finding of this study is that sigmatropic 1,5-H rearrangements proceed with H atoms within the ester groups, but not with H atoms in the polyacrylate chain. The activation energy of the latter reactions, which create ester enolates, is presumably too high. The rearrangements occurring at the ester groups are particularly useful for structural characterization of branched polyacrylates in which the branches grow out from the ester alkyl groups. These rearrangements yield fragments characteristic of the branch sizes and, in combination with multistage mass spectrometry, they also enable assessment of the branching architecture. Acknowledgements We thank the National Science Foundation for generous financial support (CHE-0517909 and DMR-0322338 and its Special Creativity Award DMR-0630301).

References 1. Tanaka K, Waki H, Ido S, Akita Y, Yoshida Y, Toshida T (1988) Rapid Commun Mass Spectrom 2:151–153 2. Karas M, Hillenkamp F (1988) Anal Chem 60:2299–2301 3. Fenn JB, Mann M, Meng CK, Wang SF, Whitehouse CM (1989) Science 246:64–71 4. Montaudo G, Lattimer RP (2002) Mass spectrometry of polymers. CRC Press, Boca Raton 5. Pasch H, Schrepp W (2003) MALDI-TOF mass spectrometry of synthetic polymers. Springer, Berlin Heidelberg New York 6. Aaserud DJ, Prokai L, Simonsick WJ Jr (1999) Anal Chem 71:4793–4799 7. Buback M, Frauendorf H, Guenzler F, Vana P (2007) Polymer 48:5590–5598 8. Norman J, Moratti SC, Slark AT, Irvine D, Jackson AT (2002) Macromolecules 35:8954–8961 9. Singha NK, Rimmer S, Klumperman B (2004) Eur Polym J 40: 159–163 10. Jackson AT, Bunn A, Priestnall IM, Borman CD, Irvine DJ (2006) Polymer 47:1044–1054 11. Bennet F, Lovestead TM, Barker PJ, Davis TP, Stenzel MH, BarnerKowollik C (2007) Macromol Rapid Commun 28:1593–1600 12. Ji H, Sakellariou G, Mays JM (2007) Macromolecules 40:3461– 3467 13. Wyzgoski F, Polce MJ, Wesdemiotis C, Arnould MA (2007) J Polym Sci A 45:2161–2171 14. Buback M, Frauendorf H, Guenzler F, Vana P (2007) J Polym Sci A 45:2453–2467 15. Jackson AT, Yates HT, Scrivens JH (1996) Rapid Commun Mass Spectrom 10:1668–1674 16. Jackson AT, Jennings KR, Scrivens JH (1997) J Am Soc Mass Spectrom 8:76–85 17. Jackson AT, Yates HT, Scrivens JH, Green MR, Bateman RH (1997) J Am Soc Mass Spectrom 8:1206–1213 18. Gidden J, Jackson AT, Scrivens JH, Bowers MT (1999) Int J Mass Spectrom 188:121–130 19. Borman CD, Jackson AT, Bunn A, Cutter AL, Irvine DJ (2000) Polymer 41:6015–6020 20. Scrivens JH, Jackson AT (2000) Int J Mass Spectrom 200:261–276 21. Jackson AT, Slade SE, Scrivens JH (2004) 238:265–277 22. Jackson AT, Williams JP, Scrivens JH (2006) Rapid Commun Mass Spectrom 20:2717–2727

Anal Bioanal Chem (2008) 392:595–607 23. Polce MJ, Ocampo M, Quirk RP Wesdemiotis C (2008) Anal Chem 80:347–354 24. Polce MJ, Ocampo M, Quirk RP, Leigh AM, Wesdemiotis C (2008) Anal Chem 80:355–362 25. Wollyung KM, Wesdemiotis C, Nagy A, Kennedy JP (2005) J Polym Sci A 43:946–958 26. Wesdemiotis C, Pingitore F, Polce MJ, Russell VM, Kim Y, Kausch CM, Connors TH, Medsker RE, Thomas RR (2006) Macromolecules 39:8369–8378 27. Chaicharoen K, Polce MJ, Wesdemiotis C (2006) Proc 54th ASMS conf mass spectrometry and allied topics. Seattle, WA, 28 May-1 June 2006 28. Matyjaszewski K (2002) Cur Org Chem 6:67–82 29. Kasko AM, Grunwald SR, Pugh C (2002) Macromolecules 35:5466–5474 30. Kasko AM, Pugh C (2004) Macromolecules 37:4993–5001 31. Pugh C, Fan G, Kasko AM (2005) Macromolecules 38:8071–8077 32. Kjellander CBK, van Ijzendoorn LJ, de Jong AM, Broer DJ, Niemantsverdriet HJW (2005) Mol Cryst Liq Cryst 434:499–510 33. Pugh C, Singh A (2007) Polym Prep 48:758–759

607 34. Singh A, Pugh C (2007) Polym Prep 48:791–792 35. Pugh C, Singh A (2007) Provisional US Patent no. 60/849,415 36. Quirk RP, Ocampo M, Polce MJ, Wesdemiotis C (2007) Macromolecules 40:2352–2360 37. Roepstorff P, Fohlman J (1984) Biomed Mass Spectrom 11:601 38. Biemann K (1988) Biomed Environ Mass Spectrom 16:99–101 39. Parsons AF (2000) An introduction to free radical chemistry. Blackwell Science, Oxford, UK 40. Benson SW (1976) Thermochemical kinetics. Wiley-Interscience, New York (Tables 2.14 and A.1) 41. Cohen N, Benson SW (1993) Chem Rev 93:2419–2438 42. Lias SG, Bartmess JE, Liebman JF, Holmes JL, Levin RD, Mallard WG (1988) J Phys Chem Ref Data 17:Suppl 1 43. http://webbook.nist.gov/ 44. Plessis C, Arzamendi G, Alberdi JM, van Herk AM, Leiza JR, Asua JM (2002) Macromol Rapid Commun 24:173–177 45. Nikitin AN, Hutchinson RA, Buback M, Hesse P (2007) Macromolecules 40:8631–8641 46. Matyjaszewski K, Gaynor SG, Müller AHE (1997) Macromolecules 30:7034–7041