Journal of Chromatography A, 1367 (2014) 90–98

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Fabrication of highly cross-linked reversed-phase monolithic columns via living radical polymerization Kun Liu a , Pankaj Aggarwal a , H. Dennis Tolley b , John S. Lawson b , Milton L. Lee a,∗ a b

Department of Chemistry and Biochemistry, Brigham Young University, Provo, UT 84602, USA Department of Statistics, Brigham Young University, Provo, UT 84602, USA

a r t i c l e

i n f o

Article history: Received 8 May 2014 Received in revised form 15 September 2014 Accepted 18 September 2014 Available online 8 October 2014 Keywords: Liquid chromatography Polymeric monoliths Capillary columns Living radical polymerization Small molecules Organotellurium

a b s t r a c t New monolithic reversed-phase liquid chromatography (RPLC) stationary phases based on single multiacrylate/methacrylate-containing monomers [i.e., 1,12-dodecanediol dimethacrylate (1,12-DoDDMA), trimethylolpropane trimethacrylate (TRIM) and pentaerythritol tetraacrylate (PETA)] were synthesized using organotellurium-mediated living radical polymerization (TERP), which was expected to produce more efficient monolithic columns than conventional free-radical polymerization. The rationale behind selection of porogens, relative concentrations of reagents and polymerization conditions are described. The new monolithic columns were applied to the separation of small molecules (i.e., alkylbenzenes) under isocratic conditions. Chromatographic efficiencies as high as 60,200 plates/m (71,300 plates/m when corrected for extra-column variance) were obtained, showing a general improvement over previous RPLC monoliths. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Organic monolithic stationary phases have made inroads as stationary phases for liquid chromatography (LC) since their introduction in the late 1980s and early 1990s [1–6]. Organic monoliths are typically prepared via conventional free radical polymerization. Compared to more conventional columns (i.e., columns packed with 5 ␮m or sub-5 ␮m particles) and silica monolithic columns, organic polymer monoliths do not perform as well for the separation of small molecules. This is primarily due to their high porosity, low mesopore volume and inherent structural heterogeneity [7,8], which result from free radical polymerization [9–13]. It is very difficult to control the free radical polymerization process and resultant monolith morphology because propagation of individual polymer chains from initiation to termination takes only seconds [14]. This characteristic of free radical polymerization usually results in the formation of microgels containing polymer chains with a heterogeneous distribution of cross-linking points [15,16] and, finally, heterogeneous polymer network structures that provide mediocre separation performance. Obviously, more homogeneous structures

∗ Corresponding author. Tel.: +1 801 422 2135. E-mail address: Milton [email protected] (M.L. Lee). http://dx.doi.org/10.1016/j.chroma.2014.09.046 0021-9673/© 2014 Elsevier B.V. All rights reserved.

with well-defined skeletal and pore sizes are desirable for obtaining better separation performance. During the past decade, controlled/living radical polymerization (CRP) was introduced and investigated for the synthesis of organic monoliths containing polar functional groups, which cannot be prepared under ionic and metal-catalyzed polymerization conditions [17–22]. Controlled/living radical polymerization is a reversible activation/deactivation process. Because of this reversible character, the growth rate of an individual chain is controlled by the balance between the growing radical and the dormant species. Therefore, the chain propagation period in CRP is much longer than in free radical polymerization, which gives the chains sufficient time to relax so that the reaction species distribute uniformly [23,24]. The most important aspect of CRP is the formation of a reversibly generated active radical. When a radical is formed, it can react with a monomer, which propagates a polymer chain and then becomes dormant [25,26]. Due to reaction reversibility, the resultant polymers exhibit narrower molecular weight distributions and more homogeneous cross-linked structures compared to polymers obtained from conventional free radical polymerization. There have been several reports of using CRP for controlling porous structures of polymer monoliths [17–22,27–32]. Yu et al. reported the use of atom transfer radical polymerization (ATRP), a CRP method, to prepare poly(ethylene glycol dimethacrylate) (PEGDMA) [33] and poly(ethylene glycol dimethacrylate-ethylene glycol methyl ether

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

methacrylate) (PEGDMA-PEGMEMA) [14] monoliths. Poly(styreneco-divinylbenzene) (PS-DVB) monoliths were successfully prepared from nitroxide-mediated living radical polymerization (NMP) [10,15,16] and reversible addition–fragmentation chain transfer (RAFT) polymerization [13,18,32]. Organotellurium-mediated living radical polymerization (TERP) is a new branch of CRP. The Yamago group investigated a series of organotellurium compounds which produce carbon-centered radicals by thermolysis to initiate polymerization reactions in the presence of an azo initiator (i.e., AIBN) [17,25,26,34,35]. This polymerization method takes place under gentle polymerization conditions, is versatile, is compatible with polar functional groups, and provides good molecular weight control [17,25,34,35]. Moreover, the reaction components involved in TERP are fewer compared to ATRP, thereby simplifying the optimization of reaction conditions. Recently, PS-DVB [11], PDVB [36], GDMA [37], and poly(N,N-methylenebisacrylamide) [12] monoliths were successfully prepared using TERP. A typical polymerization system for monolith preparation includes initiator, monomers (functional and cross-linking monomers), and porogen(s). High cross-linker concentration has been reported to provide high mechanical stability and high surface area [38,39]. Cross-linkers which contain specific functional groups provide both monolith rigidity and chromatographic selectivity [40,41]. Recent work has demonstrated that use of a singlemonomer/cross-linker in the preparation of monolithic columns provides simpler optimization of the reaction system, improved column-to-column reproducibility, better mechanical stability and higher surface area due to the highly cross-linked network [42–46]. In this study, organic monolithic capillary columns were synthesized from three different single monomers [1,12-dodecanediol dimethacrylate (1,12-DoDDMA), trimethylolpropane trimethacrylate (TRIM) and pentaerythritol tetraacrylate (PETA)] by TERP. All columns gave good separations of alkylbenzenes under isocratic conditions. This is the first report of the separation of small molecules using methacrylate/acrylate monoliths synthesized via TERP.

91

2. Experimental 2.1. Chemicals and reagents Reagents 2,2 -azobis(2-methylpropionitrile) (AIBN, 98%) and 3(trimethoxysilyl)propyl methacrylate (TPM, 98%) were purchased from Sigma–Aldrich (St. Louis, MO, USA); 1,12-dodecanediol dimethacrylate (1,12-DoDDMA), trimethylolpropane trimethacrylate (TRIM) and pentaerythritol tetraacrylate (PETA) were gifts from Sartomer (Exton, PA, USA). Ethyl-2-methyl-2-butyltellanyl propionate (BTEE; see Fig. 1 for structure) was kindly supplied by Dr. Takashi Kameshima, Otsuka Chemical Co. (Osaka, Japan). Since BTEE is oxygen sensitive, it was stored in vials that had been carefully cleaned and dried, and all transfers were conducted inside a nitrogen glove box. Water, 1,4-butanediol, propylbenzene, butylbenzene, amylbenzene and uracil were obtained from Sigma–Aldrich; acetonitrile, N,Ndimethylformamide and ethylbenzene were purchased from Fisher Scientific (Pittsburgh, PA, USA); and toluene, cyclohexanol and ethylene glycol were purchased from Mallinckrodt (Phillipsburg, NJ, USA). All solvents and chemicals for preparation of monoliths and mobile phase buffers were HPLC grade or analytical reagent grade, and they were used as received. 2.2. Fused silica capillary pretreatment First, UV-transparent fused silica capillary tubing (75-␮m, 100-␮m, and 150-␮m i.d., 375-␮m o.d., Polymicro Technologies, Phoenix, AZ, USA) was treated with TPM in order to anchor the polymer to the capillary wall. The treatment procedures were reported by Vidiˇc et al. [47] and Courtois et al. [48]. The capillary was connected to a syringe pump and washed with ethanol and then water for 30 min each. The inner surface of the capillary tubing was treated with 1 M NaOH solution at room temperature for 1 h. Both ends were then sealed with GC septa and the capillary was heated in a GC oven at 120 ◦ C for 3 h. Then, the tubing was washed with

Fig. 1. Chemical structures of BTEE and multi-methacrylate/acrylate monomers.

92

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

Table 1 Effects of reagent composition and capillary i.d. on column efficiency for poly(1,12-DoDDMA) monoliths.a , b Efficiencyc

Column Reagent composition d

Initiator

e

d

1,4-Butanediol

BTEE

19.72 18.80 17.92 16.36 15.00 13.62

0.00031 150 0.00031 150 0.00031 150 0.00031 150 0.00031 150 0.00031 150

3.0 3.0 3.0 3.0 3.0 3.0

2.8 3.0 3.2 3.6 4.0 4.5

23,400/29,900 28,600/30,100 23,500/29,900 14,700/20,700 9700/16,500

Total porogen to monomer ratio D7 0.10 28.50 53.65 D2 0.10 24.97 56.23 D8 0.10 23.21 57.50 D9 0.10 21.75 58.74 D10 0.10 19.95 60.02

17.85 18.80 19.29 19.52 20.02

0.00031 150 0.00031 150 0.00031 150 0.00031 150 0.00031 150

2.5 3.0 3.3 3.6 4.0

3.0 3.0 3.0 3.0 3.0

7700/11,000 28,600/30,100 43,800/48,000 26,600/32,400 25,400/30,500

Initiator percentage 0.05 25.02 D11 0.10 24.97 D2 0.14 25.02 D12

56.26 56.23 56.26

18.72 18.80 18.72

0.00031 150 0.00031 150 0.00031 150

3.0 3.0 3.0

3.0 3.0 3.0

26,000/27,000 28,600/30,100 10,700/20,900

BTEE volume 0.10 D13 0.10 D14 D8 0.10 D15 0.10

23.15 23.22 23.21 23.22

57.38 57.47 57.50 57.58

19.47 19.31 19.29 19.20

0.00010 150 0.00021 150 0.00031 150 0.00042 150

3.3 3.3 3.3 3.3

2.9 3.0 3.0 3.0

14,200/21,400 19,700/22,100 43,800/48,000 35,200/33,900

Capillary i.d. 0.10 D2 0.10 D16 0.10 D17

24.97 25.00 24.97

56.23 56.20 56.26

18.80 18.80 18.77

0.00031 150 0.00031 100 0.00031 75

3.0 3.0 3.0

3.0 3.0 3.0

28,600/30,100 21,600/19,000 18,500/18,100

a

c d e f

Cyclohexanol

e

Cyclohexanol to 1,4-butanediol ratio 0.10 24.97 55.31 D1 0.10 24.97 56.23 D2 0.10 24.90 57.18 D3 D4 0.10 25.00 58.64 D5 0.10 25.02 59.98 0.10 25.02 61.36 D6

b

Monomer

e

Capillary i.d. (␮m) Total porogen/monomer Cyclohexanol/1,4-butanediol Uracil/toluene

f

All monoliths were polymerized at 60 ◦ C for 24 h. Test condition: 300 nL/min flow rate. Column efficiency (plates/m). Weight percentage of monomer. Percentage by weight. Could not flush the column with 400 bar (5800 psi).

water to remove NaOH, filled with 2 M HCl solution, sealed again with GC septa and placed in a GC oven at 110 ◦ C for 3 h. The tubing was subsequently rinsed with water and ethanol and dried at 110 ◦ C with a flow of nitrogen gas overnight. Then the capillary tubing was filled with 15:85 TPM/toluene (%, w/w) solution and placed in a GC oven at 35 ◦ C overnight. Finally, the tubing was washed with toluene and acetone, successively, and dried with nitrogen gas at room temperature.

2.3. Polymeric monolith preparation All monomer solutions were prepared in 1-dram (4 mL) glass vials by admixing initiator, monomer, and porogen solvents (see Tables 1–3 for reagent compositions). All solutions were vortexed, degassed by sonication for 1 min and purged with nitrogen gas for 5 min at room temperature. The reaction promoter, BTEE, was added to the reaction solution with a 10-␮L syringe, if

Table 2 Effect of reagent composition on column efficiency for poly(TRIM) monoliths.a , b Column

Efficiencyc

Reagent composition d

Initiator

e

e

d

Cyclohexanol

1,4-Butanediol

BTEE

Total porogen/monomer

Cyclohexanol/1,4-butanediol

Uracil/toluene

Cyclohexanol to 1,4-butanediol ratio 0.10 25.03 T1 0.10 24.99 T2 0.10 24.98 T3 0.10 24.97 T4 0.10 24.85 T5

55.25 56.26 57.53 58.61 59.76

19.72 18.75 17.46 16.43 15.39

0.00031 0.00031 0.00031 0.00031 0.00031

3.0 3.0 3.0 3.0 3.0

2.8 3.0 3.3 3.6 3.9

16,100/22,900 31,200/41,300 23,800/31,200 16,700/22,800 15,100/19,900

Total porogen to monomer ratio 0.10 28.60 T6 0.10 24.99 T2 0.10 23.20 T7 0.10 21.73 T8 0.10 19.99 T9

53.52 56.26 57.60 58.72 60.00

17.89 18.75 19.20 19.55 20.01

0.00031 0.00031 0.00031 0.00031 0.00031

2.5 3.0 3.3 3.6 4.0

3.0 3.0 3.0 3.0 3.0

13,800/19,000 31,200/41,300 53,300/47,800 25,600/25,200

Control column for T7 0.10 TC

57.60

19.20

0

3.3

3.0

13,000/16,400

a b c d e f

Monomer

e

23.20

All monoliths were polymerized at 60 ◦ C for 24 h. Test conditions: 150 ␮m i.d. monolithic column, 300 nL/min flow rate. Column efficiency (plates/m). Weight percentage of monomer. Percentage by weight. Could not be measured because of large gaps in the monolith structure.

f

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

93

Table 3 Effect of reagent composition on column efficiency for poly(PETA) monoliths.a , b Column

Efficiencyc

Reagent composition d

Initiator

e

e

d

Cyclohexanol

1,4-Butanediol

BTEE

Total porogen/monomer

Cyclohexanol/1,4-butanediol

Uracil/toluene

Cyclohexanol to 1,4-butanediol ratio 0.10 25.02 P1 0.10 24.97 P2 0.10 24.97 P3 P4 0.10 24.96 P5 0.10 24.91

55.20 56.23 57.70 58.35 59.75

19.78 18.80 17.33 16.68 15.34

0.00031 0.00031 0.00031 0.00031 0.00031

3.0 3.0 3.0 3.0 3.0

2.8 3.0 3.3 3.5 3.9

31,400/35,100 39,800/58,100 39,500/46,100 39,300/42,100

Total porogen to monomer ratio P2 0.10 24.97 P6 0.10 23.21 P7 0.10 22.73 P8 0.10 21.71 P9 0.10 19.98

56.25 57.47 57.95 58.69 59.95

18.78 19.32 19.32 19.60 20.07

0.00031 0.00031 0.00031 0.00031 0.00031

3.0 3.3 3.4 3.6 4.0

3.0 3.0 3.0 3.0 3.0

39,800/58,100 43,300/60,200 40,700/42,000 24,600/29,300

a b c d e f g

Monomer

e

f

g

All monoliths were polymerized at 60 ◦ C for 24 h. Conditions: 150 ␮m i.d. monolithic column, 300 nL/min flow rate. Column efficiency (plates/m). Weight percentage of monomer. Percentage by weight. Could not flush the column with 400 bar (5800 psi). Could not be measured because of large gaps in the monolith structure.

BTEE was used. Then the reaction mixture was introduced into one end of the silanized capillary by helium gas pressure. A 5-cm long empty section was left at the other end of the capillary to provide a detection window immediately following the monolith. After filling with polymerization solution, the capillary was sealed with GC septa at both ends and heated at 60 ◦ C for 24 h. The resultant monolith was flushed with methanol and then water to remove porogens and possible unreacted residual monomers until a stable pressure reading was obtained. Monolithic columns were characterized using an FEI Helios Nanolab 600 (Hillsboro, OR, USA) scanning electron microscope after coating a thin (∼10 nm) conducting layer of gold on each capillary column end. Several SEM images were captured for the same sample at different points along the column length and analyzed using Image J software. SEM measurement of through-pore size was accomplished by measuring two orthogonal axes (longest and shortest) of the pore and taking their average as the pore diameter.

2.4. Capillary liquid chromatography An Eksigent Nano 2D LC system (Dublin, CA, USA) was used for chromatographic experiments. The injection volume was 30 nL. The two mobile phase components for elution of alkylbenzenes in RPLC were water (mobile phase A) and acetonitrile (mobile phase B). On-column detection was performed using a Crystal 100 variable wavelength UV–Vis absorbance detector (Thermo Separation Products, San Jose, CA, USA). Chrom Perfect software (Mountain View, CA, USA) was used for data collection. UV absorbance was monitored at 214 nm. 3. Results and discussion 3.1. Selection of porogens The selection of porogenic solvent(s) is an important step in the preparation of monoliths. One of the monomers, 1,12-DoDDMA,

Fig. 2. SEM images of monoliths. (A and B) Poly(1,12-DoDDMA) (D8), (C and D) poly(TRIM) (T7), (E and F) poly(PETA) (P6), and (G and H) poly(TRIM) (TC); see structures in Fig. 1.

94

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

was used to study porogen selection in this study. Several solvents with different polarities were used to synthesize the monoliths. It was found that a transparent gel or monolith was obtained when using toluene, cyclohexanol or N,N-dimethyl formamide, individually, indicating that these were potentially “good” solvents for 1,12-DoDDMA. The monomer could not be dissolved in pure ethylene glycol, 1,4-butanediol or mixtures of ethylene glycol, 1,4-butanediol and toluene. However, rigid porous monoliths were obtained when ethylene glycol or 1,4-butanediol were combined with cyclohexanol or N,N-dimethylformamide using thermal polymerization at 60 ◦ C for 24 h. Although 1,12-DoDDMA monoliths could be formed from these various combinations, those prepared from cyclohexanol and 1,4-butanediol gave lower back pressures and better chromatographic efficiencies. Consequently, cyclohexanol and 1,4-butanediol were used for preparation of poly(1,12-DoDDMA) monoliths. 3.2. Selection of polymerization conditions To obtain a homogeneous monolithic column, both gelation and phase separation must take place at the appropriate time. If gelation occurs first, the resultant structure is a gel with no distinct macroporous structure. On the other hand, early or late precipitation of the growing polymer leads to monoliths with large globular morphology and increased structural heterogeneity. The resultant monoliths have low surface area and show poor column performance for small molecules. It is desirable for gelation and phase separation to take place at the same time to produce uniform skeletal and pore structures with small pores. The effects of different polymerization reagent amounts and capillary inner diameters for poly(1,12-DoDDMA) monoliths were studied (see Table 1) to find the combination that gave the highest column efficiency. Uracil was used as a non-retained analyte, and toluene was used as a retained compound in these efficiency tests. First, the porogen ratio (i.e., cyclohexanol to 1,4-butanediol) was investigated. Other reagents and conditions were constant for all columns [i.e., total porogen/monomer was 3:1 (w/w), initiator was 0.1 wt% of monomer, promoter (BTEE) volume was 0.6 ␮L (0.00031% of the monomer amount), and capillary i.d. was 150 ␮m]. The data in Table 1 indicate that column efficiency increased with an increase in amount of cyclohexanol for uracil and toluene until the cyclohexanol to 1,4-butanediol ratio was 3:1. Since cyclohexanol is a good solvent for the growing monomer system, an increase in percentage of cyclohexanol results in smaller pore size. Moreover, an increase in cyclohexanol slows the phase separation, leading to improved monolith homogeneity because of little time difference between phase separation and gelation. Small pore size and improved monolith homogeneity lead to improved chromatographic performance of the monolithic column. Further increase in percentage of cyclohexanol leads to continuous decrease in pore size (illustrated by increase in column back pressure); however, it also delays phase separation. An extensive delay of phase separation leads to increased structural heterogeneity in the monolith and resultant decrease in column efficiency. Moreover, a point arises where gelation occurs well before the onset of phase separation, resulting in a nonporous structure as observed for column D6 in Table 1. Second, the total porogen to monomer ratio was studied for 3:1 cyclohexanol/1,4-butanediol. When the total porogen to monomer ratio was raised from 2.5:1 to 3.3:1, the column efficiency improved from approximately 7700 and 11,000 plates/m to 43,800 and 48,000 plates/m for uracil and toluene, respectively. However, with a further decrease in monomer, the monolith skeletal cross-section became thinner and the pore size became larger, as indicated by a decrease in column back pressure from 90 to 40 psi. These structural changes led to reduced column efficiency because of increased

resistance to mass transfer in the mobile phase associated with increased pore size. A decrease in column performance with an increase in pore size is well established in the literature [36]. Third, the amounts of initiator (AIBN) and promoter (BTEE) were considered. With an increase in the amount of AIBN or BTEE, the population of growing chains increased, which resulted in lower average molecular weight of each growing chain and reduction in tendency for phase separation. If the amount of AIBN or BTEE added was too small, phase separation occurred much earlier than gelation, resulting in poor column efficiency. With increasing concentration of AIBN and BTEE, column performance improved up to a certain point (Table 1) and decreased thereafter. The initial improvement in performance could be ascribed to reducing the time difference between phase separation and gelation. The eventual decrease in performance was due to extensive delay in phase separation, which resulted in a coarser bed structure and increased heterogeneity. From the experimental results, columns with an initiator percentage of 0.10% and a promoter percentage of 0.00031% BTEE gave the best performance. Finally, monolithic columns prepared in capillaries with different inner diameters (i.d.) were compared. With increasing capillary i.d. from 75 to 150 ␮m, the pore volume and column efficiency both increased. Therefore, 3:1 (w/w) total porogen to monomer ratio, 3.3:1 cyclohexanol to 1,4-butanediol porogen ratio, 0.1% initiator (i.e., 0.1% of the amount of monomer), and 0.00031% of promoter (i.e., percentage of the amount of monomer) in a 150 ␮m i.d. capillary were selected for preparation of poly(1,12-DoDDMA) monoliths. The polymerization conditions for TRIM and PETA were also investigated, as listed in Tables 2 and 3. The combination of 3:1 (w/w) total porogen to monomer ratio, and 3.3:1 cyclohexanol to 1,4-butanediol resulted in the best efficiencies for both poly(TRIM) and poly(PETA) monoliths. The effects of initiator ratio, capillary i.d. and promoter volume were not examined for either TRIM or PETA. 3.3. Monolith morphologies Fig. 2 shows SEM images of monoliths synthesized from 1,12DoDDMA (D8), TRIM (T7) and PETA (P6). From the SEM images, it appears that all three monoliths formed small globules. Poly(TRIM) had smaller through-pores than the other two monoliths, which resulted in the highest back pressure (2.25 MPa at 300 nL/min flow rate) (Fig. 2C and D). Poly(PETA) contained some large globules (Fig. 2E and F), which led to very low back pressure (0.39 MPa at 300 nL/min flow rate). 3.4. Chromatographic properties Rigid structural monoliths were obtained using all of the monomers, and all monolithic columns could be used to separate alkylbenzenes. Fig. 3 shows isocratic chromatograms of uracil, toluene, ethylbenzene, propylbenzene, butylbenzene and amylbenzene at 300 nL/min (i.e., 0.283 mm/s) with D8, T7 and P6 monolithic columns (see Tables 1–3 for column identifications). As poly(PETA) has a C5 functional group, it shows less retention and selectivity than poly(1,12-DoDDMA) and poly(TRIM), which contain more hydrophobic functionalities (i.e., C12 and C6, respectively). Poly(PETA) monoliths exhibit the weakest selectivity for alkylbenzenes; therefore, an isocratic program with less acetonitrile (40:60, v/v, water/acetonitrile) was used. As can be seen in Fig. 3, all peaks had good symmetries and narrow peak widths. The plate numbers for the columns were between 43,300 and 53,300 plates/m measured using uracil as a non-retained compound and between 47,700 and 60,200 plates/m using alkylbenzenes as retained compounds at 300 nL/min.

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

95

Fig. 4. Plate height versus linear velocity for a poly(TRIM) (T7) monolithic column using uracil as a non-retained compound () and toluene as a retained compound (䊉) (average of three repetitions). Conditions: 13 cm × 150 ␮m i.d. column; 30:70 (v/v) water/acetonitrile mobile phase; on-column UV detection at 214 nm.

Poly(1,12-DoDDMA) demonstrated the best selectivity among the three monolithic columns because it had the highest hydrophobicity compared to the other two monoliths. However, poly(PETA) and poly(TRIM) monoliths showed better column efficiencies than poly(1,12-DoDDMA), as their shorter functional group chain lengths and greater number of double bonds led to more homogeneous structures than 1,12-DoDDMA. A van Deemter curve for a T7 column was determined as shown in Fig. 4. The maximum theoretical plate number was 58,800 plates/m for uracil as a non-retained compound. This was comparable to the performance of good polymeric monoliths reported previously [40,50]. Fig. 6 shows a plot of logarithm of alkylbenzene retention factors versus carbon number. All of the dependencies in Fig. 6 are linear (R2 > 0.9965). The change in slopes for columns P6, T7 and D8 (from 0.072 to 0.1695) indicate that columns with longer hydrocarbon chain length between the methacrylate/acrylate end groups showed higher hydrophobic (methylene) selectivity.

Fig. 3. RPLC separations of alkylbenzenes on monoliths synthesized from (A) 1,12DoDDMA (D8), (B) TRIM (T7), and (C) PETA (P6). Conditions: (A) 15 cm × 150 ␮m i.d.; 30% A/70% B mobile phase; (B) 13 cm × 150 ␮m i.d.; 30% A/70% B mobile phase; (C) 12.5 cm × 150 ␮m i.d.; 40% A/60% B mobile phase; 300 nL/min flow rate; oncolumn UV detection at 214 nm. Peak identifications: uracil, toluene, ethylbenzene, propylbenzene, butylbenzene and amylbenzene in order of elution.

In recent work, we found that the injection system often contributes significant extra-column volume, which adversely affects the measured column efficiencies for small-diameter columns [49]. The extra-column volume of the injection valve for the capillary LC system used in this work was determined to be ∼15 nL. Correcting for this extra-column contribution, the column performance was found to improve by ∼57% (e.g., 53,300–84,100) for a non-retained compound (i.e., uracil) and ∼17% for a retained compound (i.e., toluene, k = 0.86). All efficiency values reported in this manuscript represent the actual measured values (unless stated otherwise) and could be corrected to indicate better actual column performance.

Fig. 5. Effect of mobile phase flow rate on column back pressure (average of three repetitions). Conditions: 10 cm × 150 ␮m i.d. monolithic columns.

96

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

Table 4 Permeabilities of poly(alkanediol multi-methacrylate/multi-acrylate) monolithic columns for different liquids. Liquid

Relative polaritya

Viscosity (mPa s)b

Water Methanol Acetonitrile

1.00 0.76 0.46

0.89 0.54 0.37

a b c d

Permeability (×10−14 m2 )c , d 1,12-DoDDMA

PETA

TRIM

3.33 ± 0.24 2.14 ± 0.22 1.71 ± 0.02

6.06 ± 0.20 4.54 ± 0.45 3.66 ± 0.04

0.82 ± 0.02 0.70 ± 0.02 0.63 ± 0.03

Relative polarity data are from Ref. [51]. Viscosity, , data are from online CRC Handbook of Chemistry and Physics, 89th ed., CRC, Boca Raton, FL, 2008–2009. Permeability k = Lu/P, where  is the viscosity, L is the column length (10 cm in this case), u is the solvent linear velocity, and P is the column back pressure. Average of six trials at different flow rates ± standard deviation.

long monolithic column at six different flow rates from 100 to 1000 nL/min. Linear relationships (Fig. 5) between back pressure and flow rate (R2 > 0.999) for all monoliths clearly indicated that the monoliths were mechanically stable. The permeabilities calculated based on Darcy’s law are listed in Table 4. All monoliths slightly swelled in acetonitrile and slightly shrunk in more polar solvents, thus leading to slightly higher water permeability. 3.6. Comparison of columns prepared from TERP and conventional free radical polymerization

Fig. 6. Plot of the logarithm of retention factors (k) of alkylbenzenes versus carbon number for the monolithic columns: () D8, () T7 and (䊉) P6. Other conditions are the same as in Fig. 3.

A control column (TC) was synthesized to compare columns prepared from the same monomer (TRIM) using TERP or conventional free radical polymerization. Comparison of the van Deemter curves for columns TC and T7 in Fig. 7 illustrate that the column prepared from TERP exhibited much lower HETP. Furthermore, the efficiency values in Table 2 clearly demonstrate an improvement in column performance for both retained and non-retained compounds. Isocratic separations of alkylbenzenes using columns TC and T7 in Fig. 8 also illustrate that the TERP column (T7) showed better selectivity for these analytes than the column synthesized

3.5. Column permeability and stability Column permeability was used to evaluate the stability of the monoliths. To obtain plots of back pressure versus flow rate, acetonitrile, methanol and water were pumped through each 10-cm

Fig. 7. Plate height versus linear velocity for poly(TRIM) monolithic columns T7 () and TC (䊉) using uracil as a non-retained compound. Conditions: () 13 cm × 150 ␮m i.d. and (䊉) 15 cm × 150 ␮m i.d. columns; 30:70 (v/v) water/acetonitrile mobile phase; on-column UV detection at 214 nm.

Fig. 8. RPLC separations of alkylbenzenes on monoliths synthesized from (A) TRIM (TC) and (B) TRIM (T7). Conditions: (A) 15 cm × 150 ␮m i.d.; 30% A/70% B mobile phase; (B) 13 cm × 150 ␮m i.d.; 30% A/70% B mobile phase; 300 nL/min flow rate; oncolumn UV detection at 214 nm. Peak identifications: uracil, toluene, ethylbenzene, propylbenzene, butylbenzene and amylbenzene in order of elution.

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

from conventional free radical polymerization (TC). In addition, the early elution of the non-retained compound on the control column in Fig. 8 indicates a very low concentration of small pores in the TC column. Comparison of the SEM images (Fig. 2C, D, G and H) also shows that the TERP column (T7) had a smaller average pore size of 3.19 ␮m in comparison to 5.63 ␮m for the column fabricated without BTEE (TC). 3.7. Reproducibility and stability of poly(TRIM) Early in this work, column reproducibility was, at best, mediocre due to the instability of BTEE. By optimizing the BTEE storage conditions (see Section 2.1), the reproducibility of the monolithic columns was improved. The run-to-run and column-to-column reproducibilities for three independent poly(TRIM) monolithic columns were measured; the RSD values based on retention times (n = 3) for run-to-run reproducibility were between 0.42% and 2.2%, and column-to-column reproducibility varied between 0.49% and 9.5%. More than 60 runs were conducted to test the robustness of the poly(TRIM) monolithic columns. There was no noticeable change observed in column performance. 4. Conclusions New monolithic RPLC stationary phases based on 1,12dodecanediol dimethacrylate (1,12-DoDDMA), trimethylolpropane trimethacrylate (TRIM) and pentaerythritol tetraacrylate (PETA) were synthesized using organotellurium-mediated living radical polymerization (TERP). These new monolithic columns gave chromatographic efficiencies as high as 60,200 plates/m for separation of small molecules (i.e., alkylbenzenes) under isocratic RP conditions. SEM images were taken which showed different globule sizes for monoliths prepared from the different monomers. Permeability tests indicated that the monoliths were mechanically stable. A comparison of a TERP column and a free radical polymerized control column demonstrated that TERP improved the morphology and separation performance of these organic polymer monolithic columns. References [1] S. Hjertén, J.-L. Liao, R. Zhang, High-performance liquid chromatography on continuous polymer beds, J. Chromatogr. 473 (1989) 273–275. [2] J.-L. Liao, R. Zhang, S. Hjertén, Continuous beds for standard and micro highperformance liquid chromatography, J. Chromatogr. 586 (1991) 21–26. [3] T.B. Tennikova, F. Svec, B.G. Belenkii, High-performance membrane chromatography. A novel method of protein separation, J. Liq. Chromatogr. 13 (1990) 63–70. [4] T.B. Tennikova, M. Bleha, F. Svec, T.V. Almazova, B.G. Belenkii, Highperformance membrane chromatography of proteins, a novel method of protein separation, J. Chromatogr. 555 (1991) 97–107. [5] F. Svec, J.M.J. Fréchet, Continuous rods of macroporous polymer as highperformance liquid chromatography separation media, Anal. Chem. 64 (1992) 820–822. [6] Q.C. Wang, F. Svec, J.M.J. Fréchet, Macroporous polymeric stationary-phase rod as continuous separation medium for reversed-phase chromatography, Anal. Chem. 65 (1993) 2243–2248. [7] I. Nischang, On the chromatographic efficiency of analytical scale column format porous polymer monoliths: interplay of morphology and nanoscale gel porosity, J. Chromatogr. A 1236 (2012) 152–163. [8] I. Nischang, I. Teasdale, O. Brüggemann, Towards porous polymer monoliths for the efficient retention-independent performance in the isocratic separation of small molecules by means of nano-liquid chromatography, J. Chromatogr. A 1217 (2010) 7514–7522. [9] K. Kanamori, J. Hasegawa, K. Nakanishi, T. Hanada, Facile synthesis of macroporous cross-linked methacrylate gels by atom transfer radical polymerization, Macromolecules 41 (2008) 7186–7193. [10] K. Kanamori, K. Nakanishi, T. Hanada, Rigid macroporous poly(divinylbenzene) monoliths with a well-defined bicontinuous morphology prepared by living radical polymerization, Adv. Mater. 18 (2006) 2407–2411. [11] J. Hasegawa, K. Kanamori, K. Nakanishi, T. Hanada, S. Yamago, Pore formation in poly(divinylbenzene) networks derived from organotellurium-mediated living radical polymerization, Macromolecules 42 (2009) 1270–1277.

97

[12] J. Hasegawa, K. Kanamori, K. Nakanishi, T. Hanada, S. Yamogo, Rigid crosslinked polyacrylamide monoliths with well-defined macropores synthesized by living radical polymerization, Macromol. Rapid Commun. 30 (2009) 986–990. [13] M. Turson, M. Zhou, P. Jiang, X. Dong, Monolithic poly(ethylhexyl methacrylateco-ethylene dimethacrylate) column with restricted access layers prepared via reversible addition–fragmentation chain transfer polymerization, J. Sep. Sci. 34 (2011) 127–134. [14] Q. Yu, Z. Qin, J. Li, S. Zhu, Diffusion-controlled atom transfer radical polymerization with crosslinking, Polym. Eng. Sci. 48 (2008) 1254–1260. [15] N. Ide, T. Fukuda, Nitroxide-controlled free-radical copolymerization of vinyl and divinyl monomers. Evaluation of pendant-vinyl reactivity, Macromolecules 30 (1997) 4268–4271. [16] N. Ide, T. Fukuda, Nitroxide-controlled free-radical copolymerization of vinyl and divinyl monomers. 2. Gelation, Macromolecules 32 (1999) 95–99. [17] S. Yamago, K. Iida, M. Nakajima, J. Yoshida, Practical protocols for organotellurium-mediated living radical polymerization by in situ generated initiators from AIBN and ditellurides, Macromolecules 36 (2003) 3793–3796. [18] M. Destarac, D. Charmot, X. Franck, S.Z. Zard, Dithiocarbamates as universal reversible addition–fragmentation chain transfer agents, Macromol. Rapid Commun. 21 (2000) 1035–1039. [19] C.J. Hawker, A.W. Bosman, E. Harth, New polymer synthesis by nitroxide mediated living radical polymerizations, Chem. Rev. 101 (2001) 3661–3688. [20] K. Matyjaszewski, J. Xia, Atom transfer radical polymerization, Chem. Rev. 101 (2001) 2921–2990. [21] M. Kamigaito, T. Ando, M. Sawamoto, Metal-catalyzed living radical polymerization, Chem. Rev. 101 (2001) 3689–3746. [22] V. Percec, A.V. Popov, E. Ramirez-Castillo, M. Monteiro, B. Barboiu, O. Weichold, A.D. Asandei, C.M. Mitchell, Aqueous room temperature metal-catalyzed living radical polymerization of vinyl chloride, J. Am. Chem. Soc. 124 (2002) 4940–4941. [23] O.W. Webster, Living polymerization methods, Science 251 (1991) 887–893. [24] G. Odian, Principles of Polymerization, 4th ed., John Wiley & Sons, Hoboken, NJ, 2004, pp. 313–330. [25] A. Goto, Y. Kwak, T. Fukuda, S. Yamago, K. Iida, M. Nakajima, J. Yoshida, Mechanism-based invention of high-speed living radical polymerization using organotellurium compounds and azo-initiators, J. Am. Chem. Soc. 125 (2003) 8720–8721. [26] S. Yamago, Development of organotellurium-mediated and organostibine mediated living radical polymerization reactions, J. Polym. Sci. A 44 (2006) 1–12. [27] C.J. Hawker, Molecular weight control by a living free-radical polymerization process, J. Am. Chem. Soc. 116 (1994) 11185–11186. [28] K. Matyjaszewski, M. Wei, J. Xia, N.E. McDermott, Controlled/living radical polymerization of styrene and methyl methacrylate catalyzed by iron complexes, Macromolecules 30 (1997) 8161–8164. [29] E.C. Peters, F. Svec, J.M.J. Fréchet, C. Viklund, K. Irgum, Control of porous properties and surface chemistry in molded porous polymer monoliths prepared by polymerization in the presence of TEMPO, Macromolecules 32 (1999) 6377–6379. [30] U. Meyer, F. Svec, J.M.J. Fréchet, C.J. Hawker, K. Irgum, Use of stable free radicals for the sequential preparation and surface grafting of functionalized macroporous monoliths, Macromolecules 33 (2000) 7769–7775. [31] C. Viklund, A. Nordström, K. Irgum, F. Svec, J.M.J. Fréchet, Preparation of porous poly(styrene-co-divinylbenzene) monoliths with controlled pore size distributions initiated by stable free radicals and their pore surface functionalization by grafting, Macromolecules 34 (2001) 4361–4369. [32] K.J. Barlow (née Tan), X. Hao, T.C. Hughes, O.E. Hutt, A. Polyzos, K.A. Turner, G. Moad, Porous, functional, poly(styrene-co-divinylbenzene) monoliths by RAFT polymerization, Polym. Chem. 5 (2014) 722–732. [33] Q. Yu, M. Zhou, Y. Ding, B. Jiang, S. Zhu, Development of networks in atom transfer radical polymerization of dimethacrylates, Polymer 48 (2007) 7058–7064. [34] S. Yamago, K. Iida, J. Yoshida, Organotellurium compounds as novel initiators for controlled/living radical polymerizations. Synthesis of functionalized polystyrenes and end-group modifications, J. Am. Chem. Soc. 124 (2002) 2874–2875. [35] S. Yamago, K. Iida, J. Yoshida, Tailored synthesis of structurally defined polymers by organotellurium-mediated living radical polymerization (TERP): synthesis of poly(meth)acrylate derivatives and their di- and triblock copolymers, J. Am. Chem. Soc. 124 (2002) 13666–13667. [36] G. Hasegawa, K. Kanamori, N. Ishizuka, K. Nakanishi, New monolithic capillary columns with well-defined macropores based on poly(styrene-codivinylbenzene), ACS Appl. Mater. Interfaces 4 (2012) 2343–2347. [37] G. Hasegawa, K. Kanamori, K. Nakanishi, S. Yamago, Fabrication of highly crosslinked methacrylate-based polymer monoliths with well-defined macropores via living radical polymerization, Polymer 52 (2011) 4644–4647. [38] B.P. Santora, M.R. Gagne, K.G. Moloy, N.S. Radu, Porogen and cross-linking effects on the surface area, pore volume distribution, and morphology of macroporous polymers obtained by bulk polymerization, Macromolecules 34 (2001) 658–661. [39] C. Viklund, F. Svec, J.M.J. Fréchet, K. Irgum, Monolithic, molded, porous materials with high flow characteristics for separations, catalysis, or solid-phase chemistry: control of porous properties during polymerization, Chem. Mater. 8 (1996) 744–750. [40] Z. Xu, L. Yang, Q. Wang, Different alkyl dimethacrylate mediated stearyl methacrylate monoliths for improving separation efficiency of typical alkylbenzenes and proteins, J. Chromatogr. A 1216 (2009) 3098–3106.

98

K. Liu et al. / J. Chromatogr. A 1367 (2014) 90–98

[41] H. Aoki, T. Kubo, T. Ikegami, N. Tanaka, K. Hosoya, D. Tokuda, N. Ishizuka, Preparation of glycerol dimethacrylate-based polymer monolith with unusual porous properties achieved via viscoelastic phase separation induced by monodisperse ultra high molecular weight poly(styrene) as a porogen, J. Chromatogr. A 1119 (2006) 66–79. [42] Y. Li, H.D. Tolley, M.L. Lee, Preparation of monoliths from single crosslinking monomers for reversed-phase capillary chromatography of small molecules, J. Chromatogr. A 1218 (2011) 1399–1408. [43] Y. Li, H.D. Tolley, M.L. Lee, Monoliths from poly(ethylene glycol) diacrylate and dimethacrylate for capillary hydrophobic interaction chromatography of proteins, J. Chromatogr. A 1217 (2010) 4934–4945. [44] K. Liu, H.D. Tolley, M.L. Lee, Highly crosslinked polymeric monoliths for reversed-phase capillary liquid chromatography of small molecules, J. Chromatogr. A 1227 (2012) 96–104. [45] C.T. Desire, R.D. Arrua, M. Talebi, N.A. Lacher, E.F. Hilder, Poly(ethylene glycol)-based monolithic capillary columns for hydrophobic interaction chromatography of immunoglobulin G subclasses and variants, J. Sep. Sci. 36 (2013) 2782–2792.

[46] R.D. Arrua, A. Nordborg, P.R. Haddad, E.F. Hilder, Monolithic cryopolymers with embedded nanoparticles. I. Capillary liquid chromatography of proteins using neutral embedded nanoparticles, J. Chromatogr. A 1273 (2013) 26–33. ˇ [47] J. Vidiˇc, A. Podgornik, A. Strancar, Effect of the glass surface modification on the strength of methacrylate monolith attachment, J. Chromatogr. A 1065 (2005) 51–58. [48] J. Courtois, M. Szumski, E. Byström, A. Iwasiewicz, A. Shchukarev, K. Irgum, A study of surface modification and anchoring techniques used in the preparation of monolithic microcolumns in fused silica capillaries, J. Sep. Sci. 29 (2006) 14–24. [49] P. Aggarwal, K. Liu, S. Sharma, J.S. Lawson, H.D. Tolley, M.L. Lee, J. Chromatogr. A (2014) (submitted for publication). ˇ ˇ [50] P. Coufal, M. Cihák, J. Suchánková, E. Tesaˇrová, Z. Bosáková, K. Stulík, Methacrylate monolithic columns of 320 ␮m I.D. for capillary liquid chromatography, J. Chromatogr. A 946 (2002) 99–106. [51] B. Gu, Z. Chen, C.D. Thulin, M.L. Lee, Efficient polymer monolith for strong cation-exchange capillary liquid chromatography of peptides, Anal. Chem. 78 (2006) 3509–3518.