MSc Chemistry
Master Track
Literature Thesis
Monolithic columns in HPLC and micro-HPLC for the
separation of small molecules
by
Nadja Volic
UvA 10628932January 2016
Supervisor:
dr. M. Camenzuli
2 | P a g e
Table of Contents
1. Abstract ... 3
2. Introduction ... 4
3. Silica-based monolithic columns ... 5
3.1. Monolithic silica rods ... 5
3.1.1. Preparation ... 5
3.1.2. Structure of monolithic silica beds ... 8
3.1.3. Chromatographic properties of monolithic silica columns ... 10
3.2. Narrow-bore monolithic columns ... 14
3.2.1. Preparation ... 14
3.2.2. The structure of monolithic silica beds in a fused silica capillary... 16
3.2.3. Chromatographic properties of narrow-bore monolithic silica columns ... 16
3.3. Second generation of monolithic columns ... 19
4. Polymer-based monoliths ... 22
4.1. Preparation ... 22
4.2. Structural characteristics ... 24
4.3. Chromatographic properties ... 26
5. Separation of small molecules on monolithic columns ... 28
6. Conclusion ... 40
3 | P a g e
1. Abstract
1. Abstract
1. Abstract
1. Abstract
Porous monolithic beds, based on organic polymers and silica, represent one of the most significant discoveries in separation science. Since their invention more than 20 years ago the interest in this type of stationary phases is in a constant rise due to their unique properties. The ease of their preparation compared to conventional packed columns, high permeability and rich chemistry (in the case of polymer-based monoliths) are reasons why monolithic columns can become an alternative to packed columns in the near future.
This review focuses on the preparation of these two types of monolithic columns and their chromatographic properties. Also application of these stationary phases in the separation of low molecular weight molecules under HPLC mode is discussed.
Keywords: Monolithic columns, Silica monoliths, Polymer monoliths, Particle packed columns, Macropores, Mesopores, Skeleton, Gel porosity, Micropores, Morphology, Permeability, Back pressure, Efficiency, Bed homogeneity, Column impedance, Small molecules.
4 | P a g e
2. Introduction
2. Introduction
2. Introduction
2. Introduction
High Performance Liquid Chromatography is nowadays the most common analytical separation tool in many industries. The number of required analysis constantly increases, requiring faster and more efficient separations.
Higher column efficiency and shorter analysis time can be obtained using smaller particles. However, the reduction in particle size leads to an increase in column back pressure (∆P∝ 1/dp2) because interstitial voids between particles become smaller. Due to existence of upper limit for the column inlet pressure (400 bar) in conventional HPLC apparatus, usage of small-sized particles imposes the application of low flow rates or short columns, in order to gain high-speed and high-efficiency separation. In conventional columns packed with 3-5 µm particles compromise has been found between back pressure and column efficiency.
In order to overcome the problem of high pressure connected with the use of small particles new approaches are developed [1-9]. The most straightforward approach to improve efficiency is to use ultrahigh pressure liquid chromatography (UHPLC) where application of high pressure enables the use of long columns packed with smaller particles. The problem of high back pressure related with the reduction in particle size can be solved by employing electroosmotic flow (EOF) in capillary electrochromatography (CEC) or by utilizing open-tube columns. Nevertheless, none of these approaches has not been employed in routine analysis due to practical difficulties [2, 6, 8].
Another way to solve this problem is to use a monolithic column that is defined as a column made of one piece of porous polymer that could be either organic or inorganic (e.g. silica) [1, 2, 3, 6, 10, 11, 12, 13]. Typical properties of this new type of stationary phase are small-sized skeleton and the presence of large through-pores (macropores) through which mobile phase flows [1, 2, 3, 10, 11]. Thin skeletons provide higher efficiency due to a reduction of diffusion path length that causes faster mass transfer [1, 2, 3, 10]. The large through-pores provide the monolithic bed with higher permeability and lower back pressure compared to the packed bed [1, 2, 3, 4, 11, 12, 14]. This enables use of high flow rates that lead to faster separation. The main advantage of the monolithic column comes from the possibility to independently control the average size of through-pores and skeleton size [1, 2, 3, 4, 10, 11, 13, 14].
5 | P a g e
3. Silica
3. Silica
3. Silica
3. Silica----based monolithic columns
based monolithic columns
based monolithic columns
based monolithic columns
At present, there are two types (which differ in inner diameter) and two generations of monolithic silica columns. The synthesis of a continuous porous silica bed is dating from 1991 [15, 16, 17, 18] while the first commercially available silica-based monolithic columns appeared on the market in 2000 [10, 12, 19]. Those columns are known under the trade name Chromolith (Merck, KGaA, Damstadt, Germany) and Onyx (Phenomenex, Torrance, CA, USA). They are both manufactured using the procedure that is patented by Merck [10].
Columns Chromolith Performance (inner diameter 4.6, 3.0 or 2.0 mm and length 100, 50 or 25 mm) and Chromolith CapRod (inner diameter 50, 100 and 200 µm and length 50,150 and 300 mm) belong to the first generation of silica-based monolithic columns [20]. Unfortunately, those columns offer limited number of bonded-phases (C18, C8, NH2 and Si).
In 2011 a new generation of monolithic columns with improved chromatographic performance appeared on the market [21]. Chromolith HighResolution columns possess inner diameter of 4.6 mm and they vary in length [22]. A few years later Kyoto Monotech launched prototype 3.0 mm inner diameter and 2.0 mm inner diameter second generation columns [21].
3.1. Monolithic silica rods
3.1. Monolithic silica rods
3.1. Monolithic silica rods
3.1. Monolithic silica rods
In this section the preparation of conventional diameter monolithic silica columns, their morphology and chromatographic properties will be discussed.
3.1.1. Prepar 3.1.1. Prepar3.1.1. Prepar 3.1.1. Preparationationation ation
The procedure developed by Nakanishi and his associates [23] represents one of the preferential approaches for the synthesis of monolithic silica rods. It consists of hydrolysis and polycondensation of a silicon alkoxide in the presence of water soluble polymer. Resulting gel is aged, dried and subsequently chemically modified and endcapped. Prior to the use, rod-shaped monolith is cladded with polymeric material. Figure 1 illustrates the procedure in which the monolithic silica column is manufactured [2].
6 | P a g e
3.1.1.1. Sol preparation and hydrolysis 3.1.1.1. Sol preparation and hydrolysis3.1.1.1. Sol preparation and hydrolysis 3.1.1.1. Sol preparation and hydrolysis
Silica monoliths are produced using sol-gel technology [10, 11, 13, 24, 25, 26, 27]. Starting materials are tetraalkoxysilanes, mostly tetrametoxysilane (TMOS) and tetraetoxysilane (TEOS) [2, 10, 11, 23, 24]. They are firstly dissolved in alcohol because of theirinsolubility in water [24]. When homogeneous solution is formed, water is added. In order to achieve hydrolysis a catalyst is needed [10, 24]. Usually 0.01 M acetic acid is used [11, 23]. During the hydrolysis alkoxy groups are replaced with silanol groups [10].The formed species condensate resulting in ether bond formation. In this way dimers, oligomers and polymers are generated.
The addition of porogen [11, 23], mostly polyethylene oxide (PEO), to the sol is needed because it forms hydrogen bridges with silanol moieties of growing polymers [2, 25]. These PEO-silica complexes are less soluble since relatively hydrophobic parts of porogen molecules become surrounded with polar solvent molecules [25]. As a result, a phase separation occurs [10].
Presence of PEO is important for one more reason. The domain size (sum of skeleton and macropore diameter) can be controlled by varying the concentration of porogen and/or tetraalkoxysilane in the synthesis mixture [1, 3, 11, 23, 28]. A decrease in PEO concentration leads to a rise in a pore and skeleton size [1, 3, 11, 23, 28] causing the formation of more homogeneous, coral-like structure (Figure 2) [28]. On the contrary, the volume of macropores depends on another parameter, concentration of solvent [2, 11, 25]. Consequently, the size and volume of macropores can be designed independently.
Figure 1. Steps involved in the preparation of a silica monolithic column for HPLC. Reproduced from [2].
7 | P a g e
3.1.1.2. Aging and drying 3.1.1.2. Aging and drying 3.1.1.2. Aging and drying 3.1.1.2. Aging and drying
The goal of the aging step is to tailor the internal porosity of the silica gel and to ensure safe drying of the monolith [10]. The aging of silica gel is performed under basic conditions. For that purpose usually aqueous solution of ammonium hydroxide is used [11, 23, 26]. Varying the concentration of ammonium hydroxide mesopores of different sizes can be obtained [1, 3].
Subsequently, the gel must be dried [10, 24]. This is a critical step because it can cause the cracking of the rod. The previous step plays a significant role because presence of small pores increases the risk of cracking [10]. During drying capillary forces play a significant role. They tend to decrease the volume of the gel.
3 33
3.1.1.3. Derivatization, endcapping and cladding .1.1.3. Derivatization, endcapping and cladding .1.1.3. Derivatization, endcapping and cladding .1.1.3. Derivatization, endcapping and cladding
The surface of the silica rod can be chemically modified in situ using octadecyldimethyl-N,N-diethylaminosilane (ODS-DEA) and endcapped with N-trimethylsilylimidazole (TMSI) [29].
During the production the silica monolith shrinks and therefore it has to be removed from a gelation tube [30] and wrapped with either polyetheretherketone (PEEK) or
polytetrafluoroethylene (PTFE) polymer [31].According to Ma [32] covering the rod with above mention materials is not sufficient because it leaves a small void between the rod and the tube
Figure 2. Scanning electron microscopy images of silica monoliths prepared byvarying the porogen concentration. Reproduced from [28].
8 | P a g e
resulting in inhomogeneous mobile phase flow through the column. As a potential solution Ma [32] and Lubda and Mullern [30] suggested alternative methods.
3.1.2. Structure of monolithic silica beds 3.1.2. Structure of monolithic silica beds3.1.2. Structure of monolithic silica beds 3.1.2. Structure of monolithic silica beds
The monolithic silica bed has a biporous structure consisting of micrometer size range through-pores (macropores) and nanometer size range mesopores (Figure 3) [4, 10, 13, 14]. Macropores form channels through which mobile phase flows while mesopores, present on the surface of silica skeleton, provide large surface area needed for retention of analytes and efficiency [10, 14].
In particle packed columns interstitial space between particles is comparable to through-pores in monolithic columns and its size strongly depends on the particle diameter (usually 25-40% of the particle size [1, 6, 8, 13]) and how densely particles are packed [1, 10, 12]. On the contrary, in monolithic columns the size of macropores and silica skeletons can be designed independently [1, 4, 10].
The volume occupied by macropores represents the external porosity of the monolithic silica column while the volume occupied by mesopores represents the internal porosity of the column [10]. The interparticle void volume represents the external porosity of packed columns [12].
The porosity of the column can be determined by inverse size exclusion chromatography (ISEC) [12, 24]. According to obtained results the external porosity of the monolithic HPLC column is 65% that is two times higher compared to columns packed with 5 µm particles (Table 1) [1, 26]. The internal porosity is moderately lower than in particle packed columns. The total
Figure 3. Scanning electron microscopy photographs of a macropore (left) and a mesopore structure of a silica skeleton (right). Reproduced from [12].
9 | P a g e
porosity (a sum of external and internal porosity) of silica monolithic columns is around 80% that is 20% higher than that found in conventional HPLC columns.
Table 1. Porosity data for a monolithic silica rod and conventional packed column. Reproduced from [26].
Volume fraction of a column
Silica rod Particle-packed
column
Pore volume Before ODS After ODS Capcellpak C18
UG
Total porosity 0.86 0.81 0.60
Through-pore 0.65 0.65 0.32
Mesopore 0.21 0.16 0.29
Bonded phase - 0.05 -
The average size of macropores and mesopores can be obtained using mercury intrusion porosimetry and nitrogen adsorption, respectively [10, 24]. Values of 2 µm for through-pores and 13 nm for mesothrough-pores are found [4, 12, 13, 14, 33]. The average size of a silica skeleton, that forms continuous bed, is estimated to be around 1 µm [1, 10]. This explains why silica monoliths have significantly higher external porosity even the size of through-pores is similar to the size of interstitial voids in conventional packed columns [1]. A surface area of 300 m2/g arises from the presence of mesopores on the surface of the stationary phase [2, 13, 14, 33, 34]. Conventional HPLC columns have a similar surface area [2 ,34].
Figure 4. Scanning electron micrograph of a monolith showing the size of macropores and silica skeletons. Reproduced from [8]
10 | P a g e
3.1.3. Chromatographic properties of monolithic silica columns 3.1.3. Chromatographic properties of monolithic silica columns3.1.3. Chromatographic properties of monolithic silica columns 3.1.3. Chromatographic properties of monolithic silica columns
In order to maintain a constant flow of mobile phase through a porous bed a pressure has to be applied at a column inlet. This pressure is defined as:
Δ =
, (1) where η is the viscosity of mobile phase, uF is the superficial velocity, L is the length of a column
(the length of a silica rod column cannot be greater than 15 cm because the column would not be straight any more [3, 7, 13]) and Kp,F is the absolute permeability.
The superficial velocity is related to the chromatographic velocity through:
uF = u0 ϵT (2)
where ϵT is the total column porosity.
The absolute permeability of a packed column is related to the external porosity (ϵe) and
particle diameter (dp) through the equation:
, =
( )
(3) while the absolute permeability of a monolithic silica column can be written as:
, =
( ) !
(4)
where l represents a parameter which value (l=2.03 µm for Chromolith columns) depends on size, size distribution, constriction and tortuosity of macropores [10].
The absolute permeability of a monolithic column can be experimentally determined by applying the equation (1). In the article published by Guiochon [10] it is reported that values for the permeability of monolithic columns that could be found in the literature are calculated using the chromatographic velocity instead of the superficial velocity. Since the former is preferred by chromatographers and the latter by engineers the chromatographic velocity will be used in this report.
In the literature it can be found that the permeability of the Chromolith column is around 8 x 10-14 m2 [9, 33, 35] which is two times higher than the value found in columns packed with 5 µm particles [9, 35]. This could be explained by much higher external porosity (equations (3) and (4)) of monolithic columns than of conventional ones. The value of 8 x 10-14 m2 corresponds to
the permeability of columns packed with 11 µm particles [33, 36, 37].
Higher external porosity of monolithic columns leads not only to the higher permeability but also to a lower back pressure under which they can operate compared to packed columns. According to Figure 5 the monolithic silica column generates the lowest back pressure under different flow rates, meaning it can perform at high flow rates allowing high-speed separation
11 | P a g e
[12]. Since commercially available monolithic columns are cladded with PEEK they cannot be used under pressure higher than 200 bar so they can operate under limited mobile phase velocities [10].
Another important characteristic of silica rod columns is high column efficiency. Two terms, plate number (N) and plate height (H), are used when column efficiency is discussed. In this report the word efficiency refers to the plate height.
Many researchers studied column efficiency of monolithic silica rod columns. The value of the optimum plate height varies from author to author. According to Minakucshi and his co-workers [1], Nakanishi and hisassociates [26] and Kele and Guiochon [38] silica rod columns exhibit minimum plate height of 12-15 µm. Additionally, Cabrera [12], Cabrera et al. [14], Kele and Guiochon [38] and Wu and co-workers [39] found plate height of around 8 µm. This difference can appear because of different methods used for determination of the column efficiency. Most widely used method uses peak width at half-height and the other method uses second moment of a peak [10, 38]. According to Kele and Guiochon [38] and Gritti and Guiochon [40] the former approach always gives lower values of H (Figure 6) because it does not take into account tailing of a peak.
Figure 5. Plot of the column back pressure against the flow rate for the monolithic column and conventional HPLC columns. Reproduced from [12].
12 | P a g e
Usually chromatographers apply the simplest form of van Deemter equation or Knox equation (equation (5) and (6)) to correlate the experimentally obtained values of H with mobile phase velocity [10].
" = # $ %$ &' ( $ &) ( (5)
" = # (+*$%
$ & (
(6)
The drawback of these approaches is that they only show to which extent each term (eddy diffusion, longitudinal diffusion, mass transfer in the mobile phase and mass transfer in the stationary phase) contributes to band broadening. They do not give any information which property of the bed causes lower efficiency.
Since the method of peak width at half-height is mainly used by researchers, values of the plate height obtained applying this method will be used to compare efficiency of monolithic columns with that of particle packed columns. Figure 7 depicts the dependence of plate height on linear velocity (van Deemter plot) for the silica rod column and columns packed with 5 and 3 µm particles [39]. From the figure it can be seen that efficiency of packed columns significantly decreases (especially for the column packed with 5 µm particles) with a rise in linear velocity contrary to the monolithic silica column. The small slope of H-u curve for the silica rod could be explained by improved mass transfer. The presence of small skeletons (the size is much smaller than the diameter of the particles) leads to the reduction of diffusion path length [1]. Also almost flat H-u curve in high velocity region means that separation can be performed at higher linear velocity (leading to fast separation) without sacrificing the efficiency.
Figure 6. HETP values obtained by using second moment method (upper curve) and half-height peak width method (lower curve). Reproduced from [38].
13 | P a g e
The reason why monolithic columns performmoderately is high eddy diffusion (A-term in equations 5 and 6) [40]. In the article by Guiochon [10] it can be found that A-term corresponds to less than one third of Hmin for particle packed columns and around two thirds for monolithic
columns. The explanation of large A-term is found in radial heterogeneity of silica rods [10, 33]. In reference [41] it can be found that the velocity of mobile phase is higher close to the wall than in the centre of the column. This means that homogeneous distribution of the macropore network is decreasing with increasing the distance from the core. The major cause of structural inhomogeneity of silica rods is the production process itself (particularly polycondensation, drying and cladding steps).
The minimum value of plate height shifts to higher values as retention factor (k) increases (equation (7)) [38]. Fornon-retained analyte (thiourea, k=0.0) optimum plate height is found to be around 7.5 µm while for triphenylene (k=4.3) the value of 13 µm is recorded.
" = 2- $ ./0$ 12(3)4
/0 ( $
1 (3)4
/5 (7) The values of retention factors obtained with monolithic silica rods are lower than those with packed columns [1, 40]. This could be explained by much greater macropore size-skeleton size ratio present in silica rods (1.2-1.5) than in packed columns (0.25-0.4) [2, 3, 5, 7, 8]. Therefore, solvents with lower elution strength have to be used [40].
In order to compare the total column performance of both packed and monolithic columns the column impedance is used [1, 10, 26]. It is defined as:
6 = 78 9 ∆; = < ,
(8) where t0 is retention time of non-retained solute, N is the number of theoretical plates, ∆P is the
column back pressure, η is the viscosity of mobile phase, H is the theoretical plate height and
Figure 7. H versus u curves of columns packed with 5 and 3 µm particles and a monolithic column. Reproduced from [39].
14 | P a g e
Kp,F is the absolute column permeability. As it can be seen from the equation (8) the column
impedance decreases as theoretical plate height decreases and as permeability increase. A study of Minakuchi et al. [1] shows that the column impedance for the column packed with 5 µm particles is around 3500 while the values for three monolithic silica columns are in the range of 800-900. More important, monolithic columns show significantly lower column impedance in high velocity area than conventional HPLC columns (Figure 8) [1, 12].
3.2. Narrow
3.2. Narrow
3.2. Narrow
3.2. Narrow----bore monolithic columns
bore monolithic columns
bore monolithic columns
bore monolithic columns
This section reports the preparation of narrow-bore monolithic silica columns, their morphology and chromatographic properties.
3.2.1. Preparation 3.2.1. Preparation3.2.1. Preparation 3.2.1. Preparation
A silica rod can be synthetized in a fused silica capillary by applying the method described in Section 3.1.1. However, there are some differences during the manufacturing process compared to previously described [5, 6, 7, 8, 42]. For instance, the silica rod is produced directly in the capillary so there is no need for cladding. The problem of shrinkage of the silica skeleton during production is also present causing the formation of voids between the rod and the inner wall of capillary. The solution for this problem is found in treating the inner wall with aqueous solution of sodium hydroxide because that would cause the attachment of the
Figure 8. Comparison of the column impedance in a monolithic column and conventional packed columns at different linear velocities. Reproduced from [12].
15 | P a g e
monolith to the wall. Unfortunately, this was a partial solution because in columns wider than 100 µm voids against the wall were still present (Figure 9) [3, 7, 8, 9].
The production of columns that have inner diameter greater than 100 µm is possible using the mixture of TMOS and methyltrimetoxysilane (MTMS) (Figure 10). Those columns are known as hybrid monolithic columns [8]. Hybrid methyl-silica monoliths are less prone to the shrinkage during synthesis and have the structure that is more flexible so it can withstand stress produced by drying of the monolith [43, 44].
Figure 9. Scanning electron microscopy photographs of a silica monolith manufactured in capillary 250 µm inner diameter column (left) and 50 µm inner diameter column (right). Reproduced from [7].
Figure 10. Scanning electron microscopy image of a hybrid monolithic column synthetized in 200 µm fused silica capillary. Reproduced from [8].
16 | P a g e
Furthermore, during the aging step urea is usually used instead of ammonium hydroxide because it simplifies production (it is added in the starting solution, which is heated, to form ammonia) and provides the monolith with more dense skeleton structure [7].
Constantin and Freitag [27] developed a single step procedure for synthesis of a monolithic silica rod in a fused silica capillary. Main advantage of their approach is the absence of volume shrinkage. The authors overcame this problem by adding dimethyldiethoxysilane in the starting solution. In this way mechanical strength was improved leading to the reduction of capillary forces that have a significant role during drying step causing a decrease of silica gel volume.
3.2.2. 3.2.2. 3.2.2.
3.2.2. The sThe sThe structure of monolithic silica beds in a fused silica capillaryThe structure of monolithic silica beds in a fused silica capillarytructure of monolithic silica beds in a fused silica capillary tructure of monolithic silica beds in a fused silica capillary
While the morphology of a silica monolith prepared in a mold can be characterized by a coral-like or spongy-like structure (Figure 2) [28], the morphology of a silica monolith prepared in a capillary can be characterized by an aggregated structure of silica globes (Figure 9) [5, 6, 7, 28]. The skeleton size is estimated to be around 2 µm while macropores are up to 8 µm large [3, 5, 6, 7, 45] which is the consequence of the skeleton attachment to the inner wall of a capillary [3, 7]. The size of the through-pores is equivalent to the interstitial space present in columns packed with 30-40 µm particles [3, 5, 6].
The narrow-bore silica column possesses the external porosity of around 86% [3, 5, 6, 28] that is 20% higher than for a monolith prepared in a mold and 50% higher than for a particle packed column (see Table 1). The internal porosity is similar to that found in monoliths synthetized in a mold [3, 7]. The total porosity is estimated to be higher than 90% [3, 5, 6, 7, 9].
3.2.3. Chromatographic properties of narrow 3.2.3. Chromatographic properties of narrow3.2.3. Chromatographic properties of narrow
3.2.3. Chromatographic properties of narrow----bore monolithic silica columnsbore monolithic silica columnsbore monolithic silica columns bore monolithic silica columns
High permeability can be ascribed to silica rods prepared in a fused silica capillary due to high external porosity (equation 4). The value of 1.3 x 10-12 m2, calculated by using the equation (1) (in this case L can be up to 150 cm [3]) can be found in the literature [7, 8, 9] meaning that narrow-bore monolithic columns have around 15 times higher permeability than silica monoliths synthetized in a mold and around 30 times higher permeability than columns packed with 5 µm particles (Section 3.1.3.).
Due to the presence of large macropores monolithic silica capillary columns generate low back pressure. As expected, conventional HPLC columns exhibit the highest back pressure while narrow-bore silica columns generate the lowest (Figure 11).
17 | P a g e
Figure 12 compares efficiency of columns packed with 5 µm particles, conventional diameter monolithic silica column and narrow-bore monolithic silica columns at different velocities. It can be seen that the column efficiency of a silica monolith synthetized in a capillary is significantly lower than that found in a silica monolith produced in a mold, especially in high velocity region. Slopes of H-u curves are steep despite the presence of small-sized skeletons [3, 5, 6]. The explanation lies in 8 µm-large macropores that cause slow mobile phase mass transfer (large Cm-term in equation (5)) [5, 6, 7, 8]. Furthermore, due to the presence of large
and small macropores (Figure 9, right one) contribution of eddy diffusion (A-term) to plate height is significant.
Figure 11. Dependence of pressure drop on velocity of mobile phase: columns packed with 5 µm particles (○,∆), silica monolith synthetizes in a mold , silica monoliths prepared in a capillary of different length (●, □, ▪, ◊, . Reproduced from [7].
Figure 12. van Deemter plots of monolithic silica columns and particle packed columns. Symbols like in Figure 11. Reproduced from [7].
18 | P a g e
Like in silica monoliths produced in a mold, the value of plate height of silica monoliths prepared in a capillary depends on the retention factor (equation (8)) [6, 7]. This is especially noticeable in high velocity region (Figure 13.)
. .
The total column performance (E, equation (8)) of the narrow-bore monolithic silica column is significantly better than that of the column packed with 5 µm particles through the whole velocity region (Figure 14) [6]. E values of the silica monolith in a capillary are slightly lower than those of the conventional diameter monolithic silica column at low linear velocities while at high linear velocities the silica rod performs better.
Figure 13. Dependence of plate height on the retention factor in narrow-bore monolithic silica columns. Symbols like in Figure 11. Reproduced from [7].
Figure 14. Comparison of the column impedance in the monolithic columns ( for capillary columns, ● for conventional diameter columns) and packed column with 5 µm particles (▪) at different linear velocities. Reproduced from [6].
19 | P a g e
3.3. Second generation of monolithic columns
3.3. Second generation of monolithic columns
3.3. Second generation of monolithic columns
3.3. Second generation of monolithic columns
High permeability and low resistance to mass transfer are two important characteristics of monolithic silica columns. The absence of radial homogeneity is the reason why they demonstrate poorer performance than one could expect. The production process itself causes a significant degree of radial inhomogeneity in silica monolithic columns [21, 41].
In order to improve chromatographic performance of monolithic silica columns some changes in the manufacturing process were made such as replacement of PEO with poly(acrylic acid) (HPAA) during gel synthesis and replacement of polymeric material (PEEK) with silicate glass in cladding step [21, 46]. Monolithic silica columns prepared in this way are known as second generation silica monolithic columns.
The usage of HPAA as a phase-separation inducer is beneficial for two reasons [21, 46]. Firstly, it eliminates radial heterogeneity. When PEO is used as the phase-separation inducer it is distributed in a silica-rich phase and it forms hydrogen bonds with silica polymers. As a result the surface of growing polymers becomes hydrophobic. Since gelation takes place in a hydrophobic mold a formation of a thick layer, so-called skin layer, close to the wall is observed causing a deformation of a forming gel in the vicinity of the wall and beneath the skin layer (Figure 15). On the contrary, HPAA is distributed in a solvent-reach phase resulting in the formation of thinner skin layer and hence less deformed silica skeleton beneath this layer (Figure 15). Secondly, the synthesis of silica rods with smaller domain size is much easier.
While the presence of 2 µm-sized through-pores and silica skeletons having a size of 1 µm is characteristic for the first generation monoliths, 1.1-1.2 µm large macropores [19, 21, 36, 47] and 0.8 µm-sized silica skeletons [21, 36] are typical for the second generation monoliths. Furthermore, the size of mesopores is shifted from 13 nm (first generation columns) to around 15 nm (new generation columns) [19, 36, 47].
Figure 16 compares the separation efficiency of the first and the second generation of monoliths. It can be seen that the HAAP columns exhibit lower plate height (H~7 µm) and that
Figure 15. Scanning electron microscopy images of monolithic silica rods prepared using HPAA (left) and PEO (right) as phase-separation inducers. Reproduced from [46].
20 | P a g e
the minimum plate height is shifted to higher linear velocity region [36, 46]. Also the slope of the H-u curve is even shallower than in the first generation of silica monolithic columns. This can be explained by the presence of thinner silica skeleton, smaller through-pores and increased radial homogeneity which leads to the reduction in eddy dispersion. On the other hand, the reduction of through-pore size causes a significant decrease in permeability meaning that second generation monoliths operates at higher back pressure. Still back pressure is in the range at which conventional HPLC apparatus operates [36, 47].
Two generations of monolithic columns differ in morphology [21], as can be seen from Figure 17. Namely, the second generation of silica monoliths is characterized by the presence of silica globules which are interconnected in a single chain, while the presence of cylindrical skeletons which form a network is typical for the first generation on monoliths.
Even though the second generation of monolithic columns shows better separation efficiency, when efficiency of more than 50 000 plates is needed, the first generation of
Figure 16. H-u curves of the first and second generation of monolithic columns. Reproduced from [36].
Figure 17. Scanning electron micrographs for the first (left) and second (right) generation of silica monolithic columns. Reproduced from [21].
21 | P a g e
monolithic columns is a better choice due to higher permeability which leads to the faster analysis [21]. Also, due to the presence of large macropores, dirty samples can be directly applied to the first generation monoliths [19, 47].
22 | P a g e
4. Polymer
4. Polymer
4. Polymer
4. Polymer----based monoliths
based monoliths
based monoliths
based monoliths
The development ofpolymeric monoliths started approximately at the same time as the development of silica monoliths. In 1989 Hjertѐn and co-workers [48] prepared a continuous compressed gel using N,N’-methylenebisacrylamide and acrylic acid. Then, Tennikova et al. published a paper on protein separation using a macroporous polymeric membrane prepared using glycidyl methacrylate and ethylene dimethacrylate [49]. Svec and Frѐchet [50] synthetized a continuous porous rod made of glycidyl methacrylate and ethylene dimethacrylate in 1992. In the following year Wang et al. [51] described the preparation of styrene-divinylbenzene-based macroporous rod.
Although researchers mainly use home-made polymeric monolithic columns there are several commercially available columns. Bio-Rad (Hercules, California, USA) offers monolithic ion-exchange analytical columns (UNO columns) made of N,N’-methylenebisacrylamide and piperizine diacrylamide [52]. Dionex (Sunnyvale, California, USA) sells monolithic ion-exchange or reversed-phase analytical columns (ProSwift columns) based on polymethacrylate and polystyrene-divinlbenzene, respectively [52, 53]. Capillary columns suitable for reversed-phase chromatography are produced by LC Packings (part of Dionex Corporation) [52].
4.1. Preparation
4.1. Preparation
4.1. Preparation
4.1. Preparation
While the production of silica-based monoliths is a multi-step process that relies on a single preparation recipe, the fabrication of organic monoliths is straightforward and is based on versatile chemistry paths resulting in their broad selectivity [54, 55, 56, 57].
Polymer-based monoliths are manufactured by in situ polymerization within a mold [55, 57, 58, 59] that can be made of different materials and have variety of formats (glass, stainless steel or PEEK tubes, fused silica capillaries, etc.) [54, 58]. A polymerization mixture comprises of monomers (functional monomer and monomer used as a cross-linker), pore-forming solvent(s) and the radical initiator [55, 57, 58]. Almost any monomer can be used in the synthesis of organic monoliths [54, 59] while the number of cross-linkers is restricted [59]. Most common cross-linkers are: divinylbenzene, ethylene dimethacrylate and N,N’-methylenebisacrylamide [59]. The choice of porogenic solvents is usually based on the experience [57, 59] and it is related to the polarity of the functional monomer and the crosslinking monomer [60]. The mixture of dodecanol (macroporogen) and cyclohexanol (microporogen) is mainly used during the synthesis of glycidiylmethacrylate/ethylene dimethacrylate-based monoliths. On the other hand, higher alcohols (macroporogens), together with toluene or THF (microporogens), are used as pore-forming solvents for the preparation of poly(styrene-co-divinylbenzene) monoliths [59, 60]. Decomposition of the radical initiator, at particular temperature, triggers the polymerization reaction [58].
The porosity of a polymeric monolith can be adjusted during the synthesis by varying the parameters such as polymerization temperature, concentration of a cross-linker, the composition
23 | P a g e
of a porogenic solvent [54, 58, 59, 60] and polymerization time [56, 59].The polymerization temperature and porogens can change porous properties of the monolith without making any changes in its chemical composition [56, 58, 59]. It is known that at lower temperatures the distribution of pore size is sharp and is shifted towards larger pores [58, 59, 60]. In addition, increasing the concentration of the macroporogen in a porogenic mixture, monolith with large pores can be obtained (Figure 18) [59, 60, 61]. Moreover, the reduction of polymerization time leads to an increase in pore volume [56, 59]. On the other hand, the concentration of a crosslinking monomer does not only have an effect on porosity, but also on a polymer composition [56, 58, 59]. Generally, the higher the concentration of a cross-linker is, the smaller the pores are [56, 58, 59, 60, 61].
Apart from thermally initiated radical polymerization, photo polymerization is the most commonly used method for the preparation of organic monoliths [55, 56, 57]. This technique is much faster than the former one, and it allows the application of porogens with low boiling points since it is performed at room temperature [56, 59, 61]. Furthermore, it leads to the formation of more radially homogeneous polymer due to the absence of temperature gradient [56]. When this technique is used to initiate polymerization, UV transparent and thin molds (glass tubes and capillaries coated with Teflon) and monomers (styrene and divinylbenzene are excluded) must be used [54, 58, 59, 61].
To initiate polymerization γ-rays or electron beam could be used, but those approaches are more unconventional [59]. Organolellirium-mediated living radical polymerization,
TEMPO-Figure 18. Scanning electron microscopy photographs of poly(glycidyl methacrylate-co-trimethylolpropane trimethacrylate) monoliths prepared by varying the ratio of a monomer and a crosslinker as well as macro- and microporogen in polymerization mixture. Reproduced from [61].
24 | P a g e
mediated living radical polymerization, atom transfer radical polymerization, ring-opening metathesis polymerization and polycondensation are also techniques that could be used during the preparation of polymeric monoliths.
In order to be suitable for a certain application, polymeric monolith has to possess particular surface chemistry [54, 55, 58, 59, 60]. The usage of monomers having certain functional groups represents the straightforward approach of making functionalized monoliths. Another option is a chemical modification of a pre-formed polymer that is synthetized using a monomer that contains a reactive group such as glycidyl methacrylate. Moreover, the surface of monoliths can be subsequently modified by attaching polymer chains to the polymeric pore surface.
4.2. Structural characteristics
4.2. Structural characteristics
4.2. Structural characteristics
4.2. Structural characteristics
The morphology of polymer-based monoliths differs significantly from morphology of silica-based counterparts (Figure 19) [56, 57, 63, 64]. While silica monoliths exhibit biporous structure, organic monoliths possess monomodal pore-size distribution [13, 37, 63]. The polymeric bed consist of nonporous microglobulus that form irregular clusters between which irregularly-shaped through-pores are generated [52, 54, 62, 64].
Figure19. Scanning electron micrographs of a silica-based monolith (left) and a polymer-silica-based monolith (right). Reproduced from [65].
25 | P a g e
Due to the absence of mesopores, polymeric monoliths have significantly lower surface area (few tens of m2/g) compared to silica-based analogues (300 m2/g) [13, 57, 62, 64]. It should be emphasized that those values are obtained by applying a dry-state method such as nitrogen adsorption. In the article of Nischang and Brüggeman [65] it is demonstrated that under chromatographic conditions porous structure of the polymer gel is different from one found in the dry state (Figure 20). When a polymer is in contact with a solvent it exhibits nano-scale gel porosity which originates from micropores (pore size less than 2 nm) [57, 64]. The degree of cross-linking of the polymer and the mobile phase composition determines the amount of gel porosity in the polymer network structure [63-67]. The higher the degree of cross-linking, the lower the contribution of gel porosity to the total polymer porosity [64]. This means that reported low surface area is not actually the surface area that is available under solvated conditions [64].
Dissimilarities in porous structure of these monoliths are the reason why their chromatographic behaviour differs significantly [56, 57, 62, 64]. Unlike silica-based monoliths, which are outstanding media for the rapid separation of small molecules under isocratic conditions, organic monoliths perform moderately under same conditions (Figure 21) [52, 56, 57, 62, 63, 64]. On the other hand, polymeric monoliths are ideal stationary phase for fast separation of biomolecules undergradient conditions [52, 56, 57, 62, 63, 64, 66, 67]. These differences in the performance of organic monoliths could be ascribed to the presence of micropores [57, 56]. Since the hydrodynamic radii of proteins are larger than the size of those pores, they are excluded from micropores and only passess macropores [65, 66]. On the other hand, separation of small molecules is affected by micropores since they enter the gel porosity region [65]. How deep small molecules will penetrate depends on their size [65, 67] and polarity [66, 67].
Figure 20. Monomodal pore size distribution of poly(butylmethacrylate-co-ethylene dimethacrylate) monolithic in the dry state and its gel porosity under solvated conditions. Reproduced from [65].
26 | P a g e
Figure 21. Isocratic separation of small molecules using Chromolith High Resolution (4.6. mm x 100 mm) column (a) and poly(styrene-divinylbenzene) monolithic column (8 mm x 100 mm) (b). Separation conditions: 80% acetonitrile in water (a) and 70% acetonitrile in water (b). Gradient elution of proteins using poly(styrene-divinylbenzene) monolithic column (8 mm x 50 mm) (c). Separation conditions: linear gradient 20-60% acetonitrile in 0.1% trifluoroacetic solution. Reproduced from [62].
4.3. Chromatographic properties
4.3. Chromatographic properties
4.3. Chromatographic properties
4.3. Chromatographic properties
In order to compare chromatographic properties of polymer-based monolithic columns with silica-based monolithic ones, commercially available polymeric column will be used in this section. In the article of Gritti and Guiochon [66], ProSwift RP-1S monolithic column (4.6 mm x 45 mm) which has macropore and skeleton sizes of approximately 1.1 µm and 2 µm, was employed. Specific surface area was estimated to be less than 2 m2 /g and the total porosity of around 57%.
Figure 22 depicts the dependence of column back pressure on the flow rate [66]. It can be seen that at lower flow rates the column back pressure is increasing steeply. According to Gritti and Guiochon [66], it is not clear why this initial and steep rise of the column back pressure occurs. Furthermore, at flow rates lower than 0.4 mL/min the permeability is 1.7 x 10-15 m2, while at higher flow rates (0.4-3.0 mL/min) it is 7.3 x 10-15 m2.
27 | P a g e
The ProSwift column exhibits significantly lower column efficiency than silica-based columns [66]. The smallest H value of around 100 µm (second moment of the peak was used to determine column efficiency) was found for non-retained compound (thiourea), while H values of retained compounds (acetophenone, valerophenone and octanophenone) were markedly higher, especially at higher flow rates (Figure 23). It can be also seen that plate height curves are steep, Hmin of thiourea arises from eddy dispersion (A-term) and that the difference in efficiency
between thiourea and three n-alkanophenones comes from C-term. Furthermore, at flow rates higher than 1 mL/min H values of octanophenone (k=12.9) are smaller than those of valerophenone (k=4.1), (equation (8)). This is probable because the former compound is more excluded from micropores than the latter one.
Figure 24 shows which sources of band broadening contribute to the plate height of small retained compounds and to which extent [66]. It can be seen that efficiency is severely limited by
Figure 22. Plot of the column back pressure against the flow rate for the
ProSwift column. Mobile phase
composition: 75% acetonitrile in water.
Figure 23. Plot of the plate height against the flow rate for non-retained and three retained compounds on ProSwift RP-1S column. Mobile phase composition: 75% acetonitrile in water. Reproduced from [66].
28 | P a g e
the very slow absorption rate (10 Hz) into the rigid, cross-linked polymer skeleton. This small frequency is due to the small specific surface area (absorption rate constant is proportional to specific surface area) and the absence of polymer swelling in the wet state. Since the monolithic polymer is highly cross-linked it stays rigid in the contact with mobile phase. Therefore, the probability that an analyte comes in contact with a soft spot, leading to analyte’s absorption, is restricted. Although the effective diffusivity of retained compounds inside the polymer skeleton is small (Ds no larger than 10% of Dm), it does not contribute significantly to the column efficiency.
The plots also show that efficiency is restricted by eddy diffusion (due to structural heterogeneity), while the contribution of longitudinal diffusion is insignificant.
5. Separation of small molecules on monolithic columns
5. Separation of small molecules on monolithic columns
5. Separation of small molecules on monolithic columns
5. Separation of small molecules on monolithic columns
Figure 24. Plots of the reduced plate height (dskel= 2 µm) against reduced velocity of
acetophenone and valerophenone reveal which sources of band broadening limit the efficiency of the ProSwift column and to which extent. Reproduced from [66].
29 | P a g e
As previously mentioned, polymer-based monolithic stationary phases are not suitable for the separation of small molecules. This is due to the high gel porosity, absence of mesopores and low surface area [68, 69, 70, 71]. Thus, several approaches have been developed in order to adjust porous property of polymer-based monoliths. These include application of higher polymerization temperature when thermal initiation is used [72], usage of lower polymerization temperature when UV initiation is applied [73], early termination of the polymerization process [68, 74, 75, 76, 77], synthesis from a single linker [71, 78, 79], usage of a longer cross-linker [69, 80, 81], postpolymerization hypercrosslinking [70, 82, 83, 84, 85, 86] and usage of carbon nanotubes and fullerenes [87, 88, 89].
Bonn’s group [68] demonstrated the effect of polymerization time (90, 180, 270 and 360 min) on the chromatographic performance of poly (N-vinylcarbazole-co-1,4-divinylbenzene) monolith. This study showed that an increase in polymerization time from 90 to 270 min leads to the change of pore size distribution (Figure 25). When polymerization reaction was terminated after 90 min a bimodal pore-size distribution was obtained and it was transferred to a monomodal pore-size distribution when a 270-min polymerization was applied. A surface area of 418 m2/g was generated after 90 min of polymerization while the lower surface area could be obtained by increasing the polymerization time.
According to Figure 26 poly(N-vinylcarbazole-co-1,4-divinylbenzene) monolith polymerized for 90 min generates significantly lower back pressure than that polymerized for 270 min, especially at higher flow rates. As it can be seen, back pressure curves are straight lines which mean that this monolith possesses high mechanical stability. Since the back pressure was lower in the case when acetonitrile was used than in the case when water was applied, it indicates that swelling propensity is low. Furthermore, the former monolith showed permeability
Figure 25. The influence of polymerization time on the pore-size distribution of poly(N-vinylcarbazole-co-1,4-divinylbenzene) monolith. Reproduced from [68].
30 | P a g e
of 1.11 x 10-14 m2 when water was used as mobile phase while the latter one showed the permeability of 4.72x 10-15 m2 (equation (1)).
The isocratic separation of seven alkylbenzens on four poly(N-vinylcarbazole-co-1,4-divinylbenzene) monolithic capillary columns (polymerization time was varying from 90 to 360 min) is shown in Figure 27. It can be seen that as polymerization time increases the separation time increases as well and hexylbenzene is not detectable anymore. The best efficiency (18500 plates/m for benzene) was obtained from the column prepared by applying polymerization time of 90 min.
Poly(N-vinylcarbazole-co-1,4-divinylbenzene) monolithic capillary column (polymerization time 90 min) demonstrated that it is efficient in the separation of small molecules (Figure 28). The lowest plate height of 3.9 µm for methylparaben (~256 000 plates/m) was obtained using this column while values of 4.8 µm (~207 000 plates/m) and 8.6 µm (~116 000 plates/m) were found for butylparabene and benzene respectively.
Figure 26. Dependence of column back pressure on the flow rate for monoliths polymerized for 90 and 270 min at different mobile phases. Reproduced from [68].
31 | P a g e
Alshitari et al. [69] examined how the length of a cross-linker affects the separation efficiency of small molecules using hexyl methacrylate-based monoliths. Two cross-linkers used
Figure 27. Isocratic separation of (1) benzene, (2) toluene, (3) ethylbenzene, (4) propylbenzene, (5) butylbenzene, (6) penthylbenzene and (7) hexylbenzene using poly(N-vinylcarbazole-co-1,4-divinylbenzene) monolithic capillary columns (0.2 mm x 80 mm). Mobile phase composition: 50% of acetonitrile in water. Flow rate: 10 µL/min. Reproduced from [68].
Figure 28. H-u curves of benzene, butylparabene and butylparabene under isocratic conditions applying poly(N- vinylcarbazole-co-1,4-divinylbenzene) monolithic capillary column (0.2 mm x 80 mm, polymerization time 90 min). Mobile phase composition: 60 % acetonitrile in water for benzene and 65% of acetonitrile in water for parabens. Reproduced from [68].
32 | P a g e
in this study were 1,6-hexanediol ethoxylate diacrylate (1,6 HEDA) and ethylene dimethacrylate (EDMA). It was found that the efficiency of poly(HMA-co-1,6 HEDA) monolith is ten times better than that of the EDMA-based monolith. This is because 1,6 HEDA has more methylene groups than EDMA which leads to higher number of mesopores.
The total porosity of poly(HMA-co-1,6 HEDA) monolith was found to be 90% leading to high permeability that was 7.75 x 10-14 m2 for acetonitrile and 1.45 x 10-13 m2 for water. Furthermore, no shrinking or swelling was found when the monolith was in contact with an organic solvent. The synthetized monoliths showed the minimum plate height of 8 µm for naphthalene using 60% acetonitrile in water as mobile phase (Figure 29).
1,6-HEDA-based monolith was used for separation of neutral non-polar compounds (alkylphenones), weak acid compounds (phenols) and basic compounds. In all three cases baseline separation was obtained (Figure 30). It is know that when basic molecules are separated on silica-based stationary phases tailing peaks are always obtained, so 1,6-HEDA-based monolith represents a promising tool for separation of small basic molecules.
.
Figure 29. H-u curve of naphthalene using poly(HMA-co-1,6 HEDA) monolith capillary column. Mobile phase composition: 60% acetonitrile in water. Reproduced from [69].
33 | P a g e
Svec and Maya [70] reported the effect of hypercrosslinking on the performance of poly(styrene-co-divinylbenzene) monoliths. In this study the hypercrosslinking was accomplished applying Friedel-Crafts reaction catalyzed by FeCl3 and using three external cross-linkers:
4,4’-bis(chloromethyl)-1,1’-byphenyl (BCMBP), α,α-dichloro-p-xylene (DCX) and formaldehyde dimethyl acetal (FDA). After hypercrosslinking monoliths had significantly larger surface area than the poly(styrene-co-divinylbenzene) monolith (Figure 31). It can be seen that the monolith hypercrosslinked with BCMBP possessed the largest surface area of 380 m2/g so it was used for testing the chromatographic performance
Figure 30. Isocratic separation of: (a) acetophenone (1), butyrophenone (2), valerophenone(3) and hexanophenone (4); (b) 4-methoxy-phenol (1), phenol (2), p-chlorophenol (3), and p-bromphenol (4); (c) quinine (1), nortriptyline (2) and amitriptyline (3); on poly(HMA-co-1,6 HEDA) monolith capillary column. Mobile phase composition: (a) 50% acetonitrile in 5 mM ammonium formate (pH=6.4); (b) 50% acetonitrile in 5 mM phosphate (pH=8) and (c) 50% acetonitrile in 5 mM ammonium bicarbonate (pH=12). Reproduced from [69].
34 | P a g e
Separation of acetone (non-retained compound) and six alkylbenzens on poly(styrene-co-divinylbenzene) and BCMBP hypercrosslinked monolithic capillary columns is shown in Figure 32. The separation of these compounds was successful when hypecrosslinked column was applied, but obtained column efficiency was not sufficient. It increased from 2900 plates/m (for benzene on poly(styrene-co-divinylbenzene) column) to 7300 plates/m (for benzene on BCMBP hypercrosslinked column).
In order to reach high efficiency separation effects such as polymerization temperature and time, amount of BCMBP, time and temperature of hypercrosslinking reaction were
Figure 31. Nitrogen adsorption/desorption isotherms determined for the monolith without hypercrosslinking and three hypercrosslinked monoliths polymerized at 5 ○C for 20h and 80 ○C for 6h, respectively. Reproduced from [70].
Figure 32. Separation of acetone (1), benzene (2), toluene (3), ethylbenzene (4), propylbenzene (5), butylbenzene (6) and n- pentylbenzene (7) applying
poly(styrene-co-divinylbenzene) (A) and BCMBP
hypercrosslinked (B) monolithic capillary columns (0.1 x 168 mm). Mobile phase composition: 60% acetonitrile and 10% tetrahydrofuran in water. Reproduced from [70].
35 | P a g e
investigated. After optimization the surface area of the hypercrosslinked monolith reached 600 m2/g and efficiency reached 71 000 plates/m for benzene (Figure 33). The column efficiency was determined under hypercrosslinking temperature and time of 90 ○C and 2 h but it did not reduce efficiency significantly.
Lee’s group [71] used polyethylene glycol diacrylate (PEGDA) of different molecular weights as a single monomer for synthesis of monoliths. Obtained results showed that monolith morphology changes from globular to fused as the molecular weight of monomer increased. Monoliths with fused morphology had lower structural heterogeneity compared to globular structure so they exhibited better column efficiency for small molecules. The column efficiency of 70 000 plates/m for uracil (non-retained analyte) was achieved using PEGDA-700 monolithic column (Figure 34).
Figure 33. H-u curve of benzene using poly(styrene-co-divinylbenzene) monolithic stationary phase hypercrosslinked with BCMBP. Conditions: polymerization temperature and time: 65 ○C and 7.5 h; hypercrosslinking temperature and time: 90 ○C and 2 h; mobile phase: 60% acetonitrile and 20% tetrahydrofuran in water. Reproduced from [70].
36 | P a g e
PEGDA-700 monolithic columns were excellent tool for separation of hydroxyl benzoic acids, phenols and alkylparabens (basic compounds). Isocratic separation of six hydroxyl benzoic acids is shown in Figure 35. It can be seen that baseline separation is achieved with column efficiency of more than 100 000 plates/m. The elution order can be explained by the presence of additional hydrogen bonding interaction apart from hydrophobic interactions. Thus, this column is suitable for the separation of polar analytes in contrast to C18 columns.
According to Figure 36 PEGDA-700 column shows excellent mechanical stability as well as little or no swelling when it is in contact with organic solvents.
Figure 34. van Deemter plot of uracile using PEGDA-700 monolithic column (column C49). Reproduced from [71].
Figure 35. Separation of benzoic acid (1), 2-hydroxy benzoic acid (2), 3-hydroxy benzoic acid (3), 3,4-dihydroxy benzoic acid (4), 3,4,5-trihydroxy benzoic acid (5), 2,4-dihydroxy benzoic acid (6) on PEGDA-700 monolithic column (column C49).
Mobile phase composition: 40%
acetonitrile with 1% of formic acid and 60% of water with 1% of formic acid. Reproduced from [71].
37 | P a g e
In contrast to monolithic columns made of polymers, silica-based monoliths have demonstrated to be excellent stationary phases for fast separation of small molecules due to the presence of mesoporous silica skeleton and surface area of 300 m2/g. The fabrication of silica monoliths is a multi-step procedure that is patented so researchers can only use commercially available columns. As Sklenáŕová et al. [19] reported, columns from a single supplier with limited number of surface chemistries like silica-based monolithic columns are not very popular. That could be one of the reasons why there is a limited number of articles about the application of monolithic columns made of silica in the literature [90, 91, 92, 93, 94, 95, 96, 97, 98, 99].
Tanaka and co-workers [90] reported the separation of five alkylbenzens in 80% methanol on C18 silica rod. According to Figure 37 baseline separation is accomplished in 30 seconds [91]. The column exhibited great performance with efficiency of 100 000 plates/m at a velocity of 5 mm/s.
Figure 36. Dependence of column back pressure on the flow rate for PEGDA-700 monolithic column (150 µm x 10 cm) at different mobile phases. Reproduced from [71].
38 | P a g e
As already mentioned, a unique characteristic of silica monolithic columns is high efficiency separation even at high flow rates without generation of high back pressure. Cabrera and his associates [14] and Majors et al. [92] demonstrated the separation of five β-blocking drugs on C18 silica rod column using different flow rates (1-9 mL/min). The drug mixture was baseline separated within 5 minutes using the flow rate of 1 mL/min (Figure 38). When applied flow rate was 9 mL/min the separation was accomplished in 30 seconds. In the former case generated back pressure was 13 bar while under the latter conditions it was 72 bar.
Figure 38. Separation of atenolol (1), pindolo (2), metoprolol (3), celiprolol (4) and bisoprolol (5) on C18 silica rod column (4.6 mm x 50 mm). Mobile phase composition: 20% acetonitrile in 0.1% trifluoroacetic acid in water. Reproduced from [14].
The separations at high flow rates can be performed on silica-based monolithic stationary phases without sacrificing the columns efficiency. This makes them suitable for flow gradient separation. The advantage of flow gradient separation over the solvent one is that no re-equilibration of monolithic column is needed at the end of each run. Previously mentioned authors [14, 92] showed the separation of 10 β-blockers using Chromolith Performance RP-18e column with and without flow gradient. As it can be seen from the Figure 39, when flow gradient is applied separation time is reduced twice. At the beginning of the separation flow rate was 2mL/min, then it changed to 5 mL/min after 3 minutes and stayed at that value for another 3 minutes.
Figure 37. Separation of benzene (1), toluene (2), ethylbenzene (3) propylbenzene (4), butylbenzene (5) and pentylbenzene (6) using C18 silica-based monolithic column (7 mm x 8.3 cm). Mobile phase composition: 80% methanol in water; linear velocity: 4.99 mm/s. Reproduced from [91].
39 | P a g e
According to Machtejevas and Cabrera [93] the application of silica monoliths in drug quality control increases. They demonstrated the separation of seven β-blockers on the second generation silica monolithic column. Figure 40 depicts the baseline separation of these compounds in 4 minutes at the flow rate of 2mL/min and back pressure of 78 bar.
Figure 39. Separation of fumaric acid (1), pindolol (2), nadolol (3), pafenolol (4), metoprolol (5), celiprolol (6), carazolol (7), bisoprolol (8), alprenolol (9) and propranolol (10) using Chromolith Performance RP-18e (4.6 mm x 50 mm). Mobile phase composition: 30% methanol in 0.02M phosphate buffer pH 3. (a) Flow rate: 2 mL/min and (b): 2 mL/min for 2 min, 2-5 mL/min in 1 min, 5 mL/min for 3 min. Reproduced from [92].
Figure 40. Chromatographic separation of atenolol (1), pindolol (2), metoprolol (3), bisoprolol (4), labetalol (5), propanolol (6) and alprenolol (7) using Chromolith High Resolution RP-18e column (4.6 x100 mm).
Mobile phase composition: 23%
acetonitrile in 20 mM potassium dihydro- phosphate buffer (pH adjusted to 2.5). Flow rate: 2 mL/min. Reproduced from [93].
40 | P a g e
6. Conclusion
6. Conclusion
6. Conclusion
6. Conclusion
A new way of producing chromatography columns represents a significant improvement in column technology that came about two decades ago. This new class of columns shows a great potential which is confirmed by a large number of published articles. The reason why their acceptance is still limited is the fact that they compete with the traditional particle-packed columns, restricted number of companies produces them and restricted surface chemistries are available in the case of silica monolithic columns.
Two main advantages of monolithic columns compared to conventional packed ones are high permeability, which is a result of the presence of large macropores, and low resistance to mass transfer due to small skeletons. Additionally, those columns are easier to manufacture and what is even more important, the size of macropores and skeleton can be adjusted independently. However, this type of stationary phases has one important shortcoming and that
41 | P a g e
is large eddy diffusion which limits the efficiency significantly. The radial inhomogeneity caused by production process itself is the reason of their mediocre performance.
The Chromolithic columns possess permeability of columns packed with 11 µm-sized particles while narrow-bore silica monolithic columns have permeability that is around 15 times higher than that of the rod columns. Due to the reduction of diffusion path length the efficiency of Chromolithic columns is comparable with that of 3.5 µm packed columns but at higher velocity region rod columns perform better. High permeability in combination with fast mass transfer enables high-speed separation to be performed on monolithic columns without loss of efficiency. The second generation of monoliths performs significantly better due to the improved production process that leads to better radial homogeneity.
The morphology of silica based monoliths differs significantly from that of polymer-based monoliths resulting in their different chromatographic behaviour. While silica-based monoliths are an ideal stationary phase for the fast separation of small molecules under isocratic conditions, polymeric monoliths are excellent media form the rapid separation of biomolecules under gradient conditions. In order to make organic monoliths suitable for the fast separation of small molecules, different approaches were developed. Some of them are application of longer cross-linker, postpolymerization hypercrosslinking, production from a single cross-linker, shorter polymerization time etc.
By reading articles about the monolithic columns one can get an impression that still a lot has to be done in this field. However, in coming years this new bed design will be competitive with particle counterparts.
7. References
7. References
7. References
7. References
1. Minakuchi, H., Nakanishi, K., Soga, N., Ishizuka, N., & Tanaka, N. (1997). Effect of skeleton size on the performance of octadecylsilylated continuous porous silica columns in reversed-phase liquid chromatography. Journal of Chromatography A, 762(1-2), 135-146.
2. Tanaka, N., Kobayashi, H., Nakanishi, K., Minakuchi, H., & Ishizuka, N. (2001). Monolithic LC columns. Analytical Chemistry, 73(15), 420A-429A.
3. Tanaka, N., Kobayashi, H., Ishizuka, N., Minakuchi, H., Nakanishi, K., Hosoya, K., et al. (2002). Monolithic silica columns for high-efficiency chromatographic separations. Journal of Chromatography A, 965(1-2), 35-49.
4. Ali, I., Gaitonde, V. D., & Aboul-Enein, H. Y. (2009). Monolithic silica stationary phases in liquid chromatography. Journal of Chromatographic Science, 47(6), 432-442.
