Optimization and development of a second-dimension separation method for the analysis of polymer nanoparticles by comprehensive two-dimensional liquid chromatography

(1)

Bachelor Thesis Scheikunde

Optimization and development of a second-dimension separation

method for the analysis of polymer nanoparticles by

comprehensive two-dimensional liquid chromatography

Door

Ruben Kers

1 juli 2016

Studentnummer

10499776

Onderzoeksinstituut Verantwoordelijke docent

Van ’t Hoff Institute for Molecular Sciences Prof. dr. ir. P.J. Schoenmakers

Onderzoeksgroep Begeleider

(2)

2

Acknowledgement

First of all, I would like to thank Bob Pirok, MSc. for his guidance and supervision during my bachelor project, his insight and willingness to share his views with me on the different problems were most supportive and certainly increased the quality of my work. Secondly, I want to thank Fleur van Beek, MSc. for being accessible and helpful with my project whenever I was in need of it. I would also like to thank prof. dr. ir. Peter Schoenmakers and dr. Wim Kok for allowing me to do my project at the Analytical Chemistry Group at the UvA. Finally, I would like to thank Serafine, Pascal, Charlotte and Leon for helping me with the practical aspects of HPLC.

(4)

4

List of Abbreviations

3-NBS acid 3-Nitrobenzenesulfonic acid

BHT Butylated hydroxytoluene

Dc Column diameter

DAD Diode array detector

e Porosity

γ Unadjusted relative retention

GFC Gel filtration chromatography

GPC Gel permeation chromatography

H Plate height

HDC Hydrodynamic chromatography

HPLC High-performance liquid chromatography

IEC Ion-exchange chromatography

L Column length

LC Liquid chromatography

LC×LC Two-dimensional liquid chromatography

N Plate number

NPLC Normal-phase liquid chromatography

PLGA Poly(lactic-co-glycolic acid)

PMMA Polymethylmethacrylate

PS Polystyrene

RID Refractive-index detector

RPLC Reversed-phase liquid chromatography

SC Slalom chromatography

SEC Size-exclusion chromatography

tR Retention time

THF Tetrahydrofuran

Ux Linear flowrate

UHPLC Ultra-high performance liquid chromatography

V0 Void volume

(5)

5

Abstract

The separation of molecules by size-exclusion chromatography (SEC) is nowadays a widely established technique in the determination of molecular weight distribution of polymers, but combining SEC with hydrodynamic chromatography (HDC) in one comprehensive analytical two-dimensional LC system (LC×LC) for the determination of polymer nanoparticles has not yet been achieved. In this study, which is part of the MAnIAC project, research is conducted towards the development of a second dimension method for this comprehensive analytical system, which is based on SEC and where HDC is applied to verify the proposed method. From this study it can be concluded that columns with high plate numbers have to be used in order to achieve good resolution when separating low-molecular weight polymers. It has been found that using multiple columns and certain core-shell columns result in sufficient separation efficiency to give good resolution between the investigated samples. HDC analysis on polymers has not been performed as method verification due to insufficient column performance.

Samenvatting

Vloeistofchromatografie is een scheidingsmethode die binnen de analytische chemie veel wordt toegepast voor de analyse van chemische stofmengsels. Bij vloeistofchromatografie wordt een scheiding uitgevoerd in een kolom, gevuld met deeltjes, waar doorheen een continue stroom van vloeistof loopt. Daarin wordt het te analyseren monster opgelost. Deze oplossing komt uiteindelijk in een detector terecht, waarbij - in het ideale geval - de individuele stoffen kunnen worden onderscheiden. Een scheidingskolom bestaat meestal uit deeltjes van vaste stof, zoals silica of koolstof. Bepaalde stoffen hebben een hogere affiniteit of passen beter door de poriën van de deeltjes van de kolom, waardoor er een scheiding wordt gedaan op basis van bijvoorbeeld hydrofobiteit, lading of grootte van moleculen.

In de analyse en scheiding van polymeren voor de bepaling van hun molmassa, wordt tegenwoordig vaak gebruik gemaakt van methoden als size exclusionchromatografie (SEC) en hydrodynamische chromatografie (HDC). Het is echter nog moeilijk om beide scheidingsmethoden in één systeem aan elkaar te koppelen. In theorie zou daardoor meer informatie verkregen kunnen worden over de te analyseren stof dan wanneer slechts één methode wordt toegepast. In het MAnIAC project wordt getracht een systeem te ontwerpen dat dit soort analyses uit kan voeren en wel met behulp van HDC en SEC.

(6)

6 In dit bachelor project ligt de focus op het vinden van een goede chromatografische methode voor het scheiden van polymeren met behulp van SEC. Hierbij wordt vooral gekeken naar welk type kolommen het beste werken voor het scheiden van polymeren met verschillende deeltjesgrootte. Het is gebleken dat een goede scheiding plaatsvindt tussen polymeren met relatief lage molmassa wanneer meerdere scheidingskolommen worden toegepast of kolommen worden gebruikt die deeltjes bevatten met een harde kern en poreuze laag. Ook is geprobeerd om HDC toe te passen voor de verificatie van de SEC-methode. Uit de resultaten is gebleken dat de HDC-kolom in onvoldoende mate functioneerde om goede analyse uit te voeren. Voor vervolgonderzoek zou dus de methode geverifieerd kunnen worden met HDC als scheidingsmethode.

(7)

7

1. Introduction

Polymers are large molecules that play key roles in everyday life and are interesting due to their wide range of properties. Polymers consist of an array of repeating subunits of carbon backbone which can contain various functional groups and heteroatoms. Examples of widely known polymers are DNA, as a naturally occurring polymer and polystyrene as a synthetically occurring polymer. Polymers are used in a broad variety of applications due to their variable chain size, hydrophobicity and functions, and are therefore incorporated in a large array of different materials.1

An important application of polymers is the synthesis of nanoparticles, which are particles with at least one dimension between 1 nm and 100 nm in size.2 An example of an application of nanoparticles is in the use of paint as coatings that have anti-reflective properties.3 These coatings are furthermore applied in for instance anti-corrosion solutions, enhanced surface appearance for furniture and UV-protection.4 For companies that produce coatings and the polymer nanoparticles they consist of, it is important to know the composition of their product and its characteristics. The knowledge on how their product behaves can be used to provide an opportunity for further development of the product. Conventionally used analytical techniques are applied to analyze these nanoparticles and polymers, such as chromatography.5

In the MAnIAC project, research is conducted towards the development of a comprehensive analytical system to determine the most important characteristics of these nanoparticles. The proposed system is portrayed in Figure 1, and is based on the principle of two-dimensional liquid chromatography (LC×LC) where two distinct separations are combined in one system for improved chromatographic results.

Figure 1: Proposed comprehensive analytical system under development in the MAnIAC

(8)

8 In the first dimension it is envisaged to separate the polymeric nanoparticles based on their size by using hydrodynamic chromatography (HDC). The nanoparticles are then to be dissolved in the modulator and the products of the dissolution are analyzed by a second dimension separation such as size-exclusion chromatography (SEC), reversed-phase liquid chromatography (RPLC) or ion-exchange chromatography (IEC).

HDC and SEC/RPLC are the methods of choice in this comprehensive system due to the important information that is acquired with these techniques. HDC is capable of separating aggregated nanoparticles in aqueous solution, while SEC is highly efficient in determining the polymers in organic solution that make up the larger nanoparticles. The proposed MAnIAC system could therefore achieve a higher selectivity compared to currently used methods, and is thus exceptionally interesting for certain applications in industry such as polymer characterization.

In this study, a possible second dimension separation mechanism is investigated on latex nanoparticles and polymers, using SEC and HDC. The results from this study can be used to fine-tune the system operating conditions like used stationary phases, flow rates, injection volumes, and dilution factors for the second dimension separation. In this research, the measurements for chromatographic optimization will be conducted using polystyrene (PS), polymethylmethacrylate (PMMA) and poly(lactic-co-glycolic acid) (PLGA) samples with differing chain lengths to measure the response of the analytical system.

Nanoparticles will be prepared with building blocks of PLGA, by making different known ratios of building blocks (consisting of PLGA and polyethylene glycol (PEG)) and separating them in the first dimension of the comprehensive system with HDC in aqueous mobile phase. The nanoparticles are then to be dissolved and separated in the second dimension with SEC in THF mobile phase, where the original building blocks can be distinguished and by which the method is verified.

It is expected that the separation of these polymers is challenging due to the wide molecular weight distribution of the analyzed molecules, resulting in broad peaks and consequently loss of resolution.

(9)

9

2. Theory

2.1 Liquid chromatography (LC)

2.1.1 The principle of LC

Liquid chromatography (LC) is a widely used separation technique which separates analytes based on their characteristics, such as hydrophobicity or molecular size. A typical LC setup consists of a pump, column and detector. In LC, the separation occurs in the column (the stationary phase) which consists of either a nonpolar material like C18, or a polar material like silica for respectively reversed-phase liquid chromatography (RPLC) or normal-phase liquid chromatography. Another mode of separation in LC is based on the molecular size of the analytes, which is the separation mechanism that occurs in size-exclusion chromatography (SEC) and hydrodynamic chromatography (HDC). The analyte is dissolved in the mobile phase, which usually consists of either acetonitrile, methanol, water, THF or mixtures of solvent that flows continuously through the system.6 In Figure 2 a typical setup of an LC system is shown.

Figure 2: Schematic of an LC setup.7

When pressures of up to 400 bar are applied for separation in LC, the technique is called high-performance liquid chromatography (HPLC). When operating under ultra-high pressures (~1000 bar) the technique is referred to as ultra-high performance liquid chromatography (UHPLC). The benefit of using high pressure is the fact that separation can be performed faster, and smaller particles in the stationary phase can be used which increases resolution at the expense of increased backpressure.8

(10)

10 LC is nowadays used as a method of analysis or as a purification method, due to the fractionation that occurs in the column.6 The LC output can be subsequently coupled to different types of detection methods, like UV detection, refractive-index detection, fluorescence detection, evaporate light scattering detection or mass spectrometry (optical rotation detection, tandem mass spectrometry).6_{In this study, HPLC is applied with UV and} refractive-index detection.

2.1.2 Resolution and band broadening

The resolution of separation in LC depends on two factors: the difference in elution time of compounds and the band broadening of peaks. There are multiple approaches to influencing the resolution, for example by changing the elution strength of the mobile phase, changing the temperature of the column, changing particle size of the column and changing the column length.6 In general, the higher the amount of theoretical plates, the better the resolution (see Eq. (1)):

𝑅𝑒𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 √𝑁

 𝛾 −  (1)

Where N is the number of theoretical plates, and γ is the unadjusted relative retention.6

The amount of plates on a column is related to the length of the column and the plate height by Eq. (2):

𝑁𝐿

𝐻 (2)

Where N is the number of plates, L is the column length and H is the plate height.6

The influence of the flow rate on the plate number and plate height in a certain column can furthermore be described by the Van Deemter Equation (Eq. (3)):

𝐻 ≈ 𝐴  𝐵

𝑈𝑥 𝐶𝑈𝑥 (3)

Where H is the plate height, A, B and C are constants dependent on the multiple flow paths within the column, longitudinal diffusion and equilibration time, respectively. Ux is the linear velocity.6

(11)

11 The Van Deemter Curve describes the column efficiency of the stationary phase and is dependent on the band broadening processes that occur during separation. Band broadening occurs for instance due to Eddy diffusion, which is related to the multiple flow paths that molecules take when going through the column. Flowrates are different in narrow and broad flow paths, which causes different velocities in the same column and results in band broadening. The A-term in the Van Deemter curve describes this Eddy diffusion. Furthermore, longitudinal diffusion causes band broadening due to dispersion of analyte within the column, which is described by the B-term in the Van Deemter Curve. Also the resistance of the analyte to mobile-phase mass transfer is an important cause of band broadening, which means that solute near the particles of the packing move more slowly through the column than the solute that is further away from the packing. This is described by the C-term of the Van Deemter curve. The characteristics of the column packing, like particle size and pore sizes have significant influence on band broadening processes. To optimize resolution, it is therefore important to take into account the type of stationary phase used.9

The resolution of separations, plate number and plate height are important factors in the development of useful LC×LC analysis methods. However, the magnitude of these effects are often times best to be determined experimentally and therefore the aim of this study is to find optimal conditions in which good polymer separations can be performed.

2.2 Size-exclusion chromatography (SEC)

2.2.1 The principle of SEC

Size exclusion chromatography (SEC) is a variant of chromatography in which compounds are separated based on their size. The stationary phase consists of porous packing material, in which small particles diffuse through pores of the column packing, whereas larger molecules are excluded of the pores. There is no interaction with the stationary phase other than the selective diffusion through the pores that is caused by entropic effects.9 The result is a separation based on the difference in rate at which molecules migrate through pores of the stationary phase particles, which depends on the molecular weight and hydrodynamic size of the molecules.5, 10 This process of size exclusion during a SEC separation is illustrated in Figure

(12)

12

Figure 3: Illustrative schematic of exclusion of compounds within pores in SEC.

SEC is often applied for the determination of the average molecular weight of biopolymers and synthetic polymers. There are various types of SEC columns, they differ in for example pore size, column length and column diameter. In gel permeation chromatography (GPC), a stationary phase consists of a hydrophobic gel in combination with an apolar mobile phase, while in gel filtration chromatography (GFC), a hydrophilic packing material is used in combination with a polar mobile phase. Packing material with distinct pore sizes are produced for optimal separation of different polymer size ranges and there are multiple types of columns available for a broad series of polymers.11

2.2.2 Practical aspects of SEC

A typical size determination measurement in SEC starts with the setup of a calibration curve with known molecular weight standards. The logarithm of the molar weight is plotted against the retention volume (VR), from which the molecular weight of the unknown sample is calculated. A calibration curve is different for every system, due to the multiple factors involved that determine retention in a column. Parameters which are of significant influence on the measurements are for example the size of the packing material, the material of which the packing material consists, the inner diameter and the length of the used columns.9 Furthermore, the hydrodynamic size of the analyzed polymers has to be taken into account and the right polymer standards need to be used, as polymers have different ways in which they are folded.5 Calibration measurements are applied in this study for the calculation of polymer molar mass, and to confirm that the utilized SEC method provides good results with known PLGA samples. An example of a SEC calibration curve is portrayed in Figure 4.

(13)

13

Figure 4: Example of a SEC calibration curve of polystyrene.12

The maximum size of molecules that can be separated on a SEC column elute on the left side of the calibration curve that is following an increasing vertical trend, which is termed the exclusion limit of the column and is caused by the fact that those molecules are of sizes too large to enter the pores of the stationary phase and thus go through the column without being retained. The other end of the calibration curve is called the permeation limit, which is achieved when molecules have such a low molar mass that they penetrate the pores of the stationary phase completely. The t0 marker, which in this study is toluene in SEC, elutes at the permeation limit.11 The V0 and therefore t0 in SEC can be predicted theoretically with Eq. (4):

𝑉  𝐷𝑐𝐿𝑒



𝜋 (4)

Where V0 is the void volume, dc is the column diameter, L is the length of the column and e is the porosity.13

When molecules reach a certain hydrodynamic size above the exclusion limit of a column, they can start to unfold and stretch within the stationary phase, which causes a phenomenon that is called slalom chromatography (SC). SC is caused by the stretching of molecules turning around the stationary phase packing material, which results in increased retention and is an entirely different retention mechanism than that of SEC. The tendency in which SC occurs in the stationary phase is dependent on the size of the analyzed molecules, the column packing, the flow rate, solvent viscosity and temperature of the column.9 In this study, SC is observed in the calibration of columns with polymer standards, but not during PLGA analysis due to the small molecular weight of the samples used in this study.

(14)

14 Current SEC methods show promising results in determining the characteristics of polymers, and it is therefore an increasingly important technique in the determination of molar weight of particles in the 100 Å to 10.000 Å linear dimension range.10 Furthermore, one of the features of polymer analysis with SEC is that the molecular weight distribution can be determined, due to the differing polydispersity of the polymers and elution with a Gaussian distribution. Both biopolymers and synthetic polymers diverge from narrowly to broadly polydisperse, which results in varying broadness of peaks on the chromatogram. Due to the wide distribution of the molar weight, there is potential loss of resolution which poses a challenge in choosing the right stationary phases in SEC.9 Figure 5 depicts a mass spectrum of polystyrene, showing the distribution of polymer molecular weight caused by the polydispersity, resulting in broad peaks in SEC chromatograms.

Figure 5: MALDI-ToF-MS spectrum of polystyrene with a molecular weight of 5050 g/mol and

(15)

15 2.2.3 Core-shell columns

An emerging new technology in HPLC is the application of core-shell particles in SEC separations. Core-shell particles are particles of porous silica or C18 material which is wrapped around a solid core, resulting in a particle size of ~2.6 μm, but which has the efficiency of a sub-2 μm particle size column. This increased efficiency is induced by the lower pore volume in core-shell columns, which lowers the longitudinal diffusion in the column, but also decreases the A- and the C-term in the Van Deemter curve due to faster mass transfer and rougher surface of the particles. The disadvantage of core-shell columns is the fact that the pore volume is significantly smaller than for columns with completely porous particles, which retarded the use of core-shell columns due to fear of lower separation capability. But on the contrary, the latest studies on core-shell columns show that these stationary phases are readily capable of excellent resolution and fast separations on UHPLC devices.

In this study, experimental Phenomenex silica and C18 core-shell columns will be applied in the separation of polymers, to investigate if these core-shell columns are indeed capable of increased resolution in PLGA separations.

2.3 Hydrodynamic Chromatography (HDC)

2.3.1 Principle of HDC

A chromatographic method which is exceptionally useful in the determination of nanosize particles is HDC, in which non-porous inert particles in packed columns or open tube capillary columns are applied to induce separation, and in which the size, the shape and the structure of analytes can be determined, when coupled to a variety of detection methods. The separation in HDC is based on the Poiseuille-like (parabolic) flow profile, which is the result of different flow lines within the column. Flow lines that are closer to the walls of the column or to the stationary phase particles move slower than flow lines in the middle of the column. Due to the capability of smaller molecules to approach the walls or particles of the column more closely, they elute at a later time compared to the larger particles that experience a greater net flow.10

(16)

16

Figure 6: Mechanism of HDC separation in an open tubular column and a packed column.10

HDC is originally an early established technique in chromatography, but has undergone a resurgence because of its capability in separating molecules with higher molar mass than SEC.10_{The use of HDC columns is attractive when large hydrophobic polymers are being} analyzed in aqueous samples, due to micelle formation which is the result of the aggregation of hydrophobic material in order to reduce its contact area with the hydrophilic environment. The micelles itself are larger particles than the un-aggregated polymers, scaling up their hydrodynamic size into the nanoparticle range and thus making HDC an excellent technique for separation.10Figure 7 portrays a schematic of a polymer micelle.

(17)

17 2.3.2 Practical aspects of HDC

In HDC, packed columns are more conventionally used due to the narrow channel that is necessary in open HDC columns. Using open HDC columns is difficult, which is especially caused by the requirement of low injection volumes (~1 µL) to avoid column overloading.5

HDC effects arise during separations of very large molecules and for columns with small pore sizes, which is a widely occurring phenomenon due to the fact that at a certain point of molecular weight the molecules approach the exclusion limit and will stop transfering through the porous stationary phase. It is therefore possible to perform HDC with size-exclusion columns. The rate at which HDC effects occur depends on the ratio of the pore size of the particles to the diameter of the particles (Rp/dp), a decrease in this ratio results in less

HDC effects in the stationary phase and a more dominant SEC separation mechanism during the separation.9 For instance in this study, HDC effects are induced with experimental SEC columns that contain small porous particles, on molecules with high molecular weight.

In the determination of molar weight distribution of polymers with HDC, it is standard procedure to first construct a calibration. A calibration is performed by injecting polymers of known sizes and measuring the retention time of the polymer with corresponding molar weight, just like in SEC. An illustration of different calibration curve regions for HDC, SEC and SC is given below in Figure 8.17

Figure 8: Calibration with polystyrene showing the specific molar weight regions in which

(18)

18

3. Experimental Section

3.1 Chemicals

Tetrahydrofuran (THF, AR-grade, stabilized and HPLC-S grade, unstabilized) was obtained from Biosolve B.V. (Valkenswaard, the Netherlands) and used for the preparation of samples and as SEC mobile phase. Butylated hydroxytoluene (BHT) was obtained from Sigma Aldrich (St. Louis, MO, USA) and was applied as HPLC-S grade THF stabilizer. Toluene was acquired from Biosolve B.V. (Valkenswaard, the Netherlands) and was applied as SEC marker. Deionised water (Arium 611UV, Sartorius, R=18.2 MΩ cm-1_{, Germany) was prepared as HDC} mobile phase and applied in sample dilutions. HDC buffer was obtained from Agilent Technologies (Amstelveen, the Netherlands) and used as HDC mobile phase surfactant. Potassium dichromate was purchased from Sigma Aldrich (St. Louis, MO, USA), uracil was obtained from Sigma Aldrich (St. Louis, MO, USA) and 3-nitrobenzenesulfonic acid (3-NBS acid) was received from Sigma Aldrich (St. Louis, MO, USA), which were used as an HDC marker. Polystyrene (PS) and poly(methylmethacrylate) (PMMA) samples with different polymer length were obtained from Polymer Laboratories (Amherst, U.S.). Poly(lactic-co-glycolic acid) (PLGA) samples with a molar weight distribution of 21 kDa and 10.5 kDa and large polydispersity PS with molecular weight distribution of 200 kDa were received from DSM Coating Resins (Waalwijk, the Netherlands).

3.2 Instrumental

All HPLC measurements using RID detection were conducted on a combined Agilent 1100 series and 1260 series LC setup, which consisted of an 1100 series nanopump (G2226A), an 1100 series refractive-index detector (G1362A), an Infinity 1260 series degasser (G1322A) and an 1100 series thermostatted column compartment (G1316A). Manual injection was used with a 20 μL sample loop installed.

All HPLC measurements using UV detection were conducted on a combined Agilent 1100 series, 1260 series and 1290 series LC setup, which consisted of a 1100 series quaternary pump (G1311A), a 1290 series diode array detector (G4212A), an Infinity 1260 series degasser (G1322A), and an 1100 series thermostatted column compartment (G1316A). An autosampler was used with 20 µL injections, consisting of an 1100 series ALS (G1313A).

(19)

19 All UHPLC measurements were conducted on an Agilent 1290 Infinity LC system, which consisted of an Infinity 1290 binary pump (G4220A), an Infinity 1290 diode-array detector (G4212A), an Infinity 1290 autosampler and an Infinity 1290 thermostatted column compartment (G1316C).

3.3 Analytical Methods

3.3.1 Sample Preparation

SEC calibration standards were prepared separately by dissolving 100 mg of either PS or PMMA with known molecular weight in 5 mL of THF, to a concentration of 20 mg/mL. The SEC calibration stock solutions were subsequently diluted in THF to a concentration of 0.5 mg/mL. The molecular weight of both used PS and PMMA standards are given on the next page in Table 1.

(20)

20

Table 1: Molecular weight of the used PS and PMMA SEC calibration standards.

PS Standard Molar Weight (g/mol) PMMA Standard Molar Weight (g/mol) PS01 580 PMMA01 620 PS02 980 PMMA02 1140 PS03 1990 PMMA03 1310 PS09 9920 PMMA04 1960 PS10 13880 PMMA05 2000 PS11 19880 PMMA06 2870 PS12 52400 PMMA07 6820 PS13 70950 PMMA08 10260 PS16 299400 PMMA09 14920 PS19 735500 PMMA10 24830 PS21 1373000 PMMA11 30690 PS22 2061000 PMMA12 49600 PS23 2536000 PMMA13 79500 PS24 3053000 PMMA14 100000 PS25 3743000 PMMA15 141500 PS26 7450000 PMMA16 211400 PS27 13000000 PMMA17 300300 - - _{PMMA18 518900} - - _{PMMA19 659000} - - _{PMMA20 948500} - - PMMA21 1250000

PLGA samples for SEC were prepared by dissolving 64.3 mg of (PLGA(7.5 kDa))2-PEG(6 kDa) (PLGA 21 kDa) in 1 mL THF, 63.4 mg (PLGA(3.75kDa))2-PEG(3 kDa) (PLGA 10.5 kDa) in 1 mL THF and dissolving 30.2 mg PLGA (21 kDa) combined with 31.1 mg PLGA (10.5 kDa) in 1 mL THF. The samples were subsequently diluted 10, 50 and 100 times in THF to a volume of 1 mL.

(21)

21 A vial of 3-Nitrobenzenesulfonic acid (3-NBS acid) was prepared to a concentration of 2.0 mg/mL in HDC buffer. 301 mg Potassium dichromate was dissolved in 5 mL HDC buffer after which the solution was diluted to a concentration of 0.301 mg/mL in HDC buffer. 200 mg Uracil was dissolved in 5 mL warm HDC buffer and diluted to a concentration of 2.0 mg/mL. The prepared solutions were used as HDC t0 marker.

3.3.2 Separation of PLGA with different SEC columns

To determine whether the PLGA particles could efficiently be separated with the available PLGel columns, different PLGA sample mixtures were injected into the HPLC system using PLGel Mixed-E (7.5x100 mm, 3 μm), PLGel Mixed-D (7.5x300 mm, 5 μm) and PLGel Mixed (7.5x600 mm) columns, using refractive-index detection. The flowrate was set to 1.0 mL/min. The mobile phase consisted of 100% THF (AR-grade, stabilized). The column compartment and the RID had a temperature set to 30 °C. The injection volume was 20 μL for every measurement.

To test the efficiency of a Phenomenex core-shell B109 column (4.6x150 mm, 2.6 μm, XB-C18), the different PLGA sample mixtures were injected into the HPLC system using refractive-index detection. The flowrate was set to 0.2 mL/min. The mobile phase consisted of 100% THF (HPLC-S grade, stabilized). The column compartment and the RID had a temperature set to 30 °C. The injection volume was 20 µL for every measurement.

3.3.3 Calibration of size-exclusion columns

Calibrations with PS and PMMA were performed by injecting every PS and PMMA standard (see Table 1) into the HPLC setup with RID detection. A PMMA and PS calibration was conducted with PLGel Mixed (7.5x600 mm), with a flowrate of 1.0 mL/min for every measurement. The mobile phase consisted of 100% THF (AR-grade, Stabilized). The used column and RID temperature was set to 30 °C. The injection volume was 20 µL for every measurement.

A calibration was performed on a Phenomenex core-shell B109 column (4.6x150 mm, 2.6 μm, XB-C18), by injecting PMMA standards (see Table 1) with the HPLC setup on RID detection. A flowrate of 0.2 mL/min was used for every measurement. The mobile phase consisted of 100% THF (HPLC-S grade, unstabilized) with added BHT (to 200 ppm). The used column and RID temperature was set to 30 °C. The injection volume was 20 µL for every measurement.

(22)

22 A calibration with PS (see Table 1) was performed on three coupled Phenomenex core-shell B111 (4.6x150 mm, 3.6 μm, XB-C18), B104 (4.6x150 mm, 2.6 μm, silica) and B109 (4.6x150 mm, 2.6 μm, XB-C18) columns, on an UHPLC system with UV detection. The flowrate was set to 2.00 mL/min for every measurement. The mobile phase consisted of 100% THF (HPLC-S grade, unstabilized) with added BHT (to 200 ppm). The column compartment had a temperature set to 25 °C. The injection volume was 1 µL for every measurement.

3.3.4 Analysis of large polydispersity PS

To see how the Phenomenex core-shell B111, B104 and B109 columns respond to large polydispersity polymers, a large polydispersity PS sample with a molecular weight of 200.000 g/mol was injected on an UHPLC system with UV detection. The flowrate was set to 2.00 mL/min for every measurement. The mobile phase consisted of 100% THF (HPLC-S grade, unstabilized) with added BHT (to 200 ppm). The column compartment had a temperature set to 25 °C. The injection volume was 1 µL for every measurement.

3.3.4 HDC column testing

An injection with 3-NBS acid and uracil was performed using an Agilent PL-PSDA Cartridge Type 1 HDC column (7.5x800 mm), on an HPLC system with UV detection. The flowrate was set to 1.6 mL/min for the 3-NBS acid injection, uracil injection and a blank sample, containing acidified HDC buffer (pH ~3.2). The mobile phase consisted of acidified HDC buffer. The used column temperature was set to 25 °C. The injection volume was 0.5 μL for 3-NBS acid and 5 μL for uracil.

An injection with potassium dichromate was performed using an Agilent PL-PSDA Cartridge Type 1 HDC column (7.5x800 mm), on an HPLC system with UV detection. The flowrate was set to 1.7 mL/min for both the potassium dichromate injection and the blank sample, containing HDC buffer (pH ~7). The mobile phase consisted of HDC buffer. The used column temperature was set to 25 °C. The injection volume was 10 μL.

(23)

23

4. Results & Discussion

4.1 Separation of PLGA with PLGel columns

4.1.1 Column efficiency determination with PLGA

Figure 9: Chromatogram of PLGA mixture, blank THF sample and toluene marker, separated

with a PLGel Mixed-E column.

Figure 9 shows the chromatogram of the PLGA mixture sample separated on a PLGel

Mixed-E column. As illustrated in the figure, there is a signal visible at 1.99 min. which is assigned to the PLGA sample, it is absent in the blank sample and elutes before the toluene marker, which elutes at 3.09 min. The t0 marker has a retention time that sufficiently corresponds to the theoretically predicted value of 3.27 min., which is the retention time at the permeation limit and is calculated according to Eq. (4). The positive signal at 2.73 min. and the strong negative signals are suspected system peaks. The positive system peaks are most likely the result of components present in the sample but absent in the eluent, which could be due to either contamination of the sample or the formation of peroxides in the sample that form when the THF reacts with oxygen. The negative signals are caused by compounds in the eluent that are not present in the injection, which could for example be peroxides that exist in a different concentration than in the sample.18_{Figure 10 depicts the chromatogram of this PLGA sample.}

-8000 -6000 -4000 -2000 0 2000 4000 6000 0 1 2 3 4 5 6 Sig n al (n RIU ) Time (min)

PLGel Mixed-E Column

Blank t0 PLGA

(24)

24

Figure 10: Chromatogram of PLGA mixture, separated with a PLGel Mixed-E column.

There is no separation between both PLGA samples visible from this chromatogram, but a broad distribution is observed which is assumed to correspond to both PLGA samples. From this result it can be concluded that the used column was not capable of giving sufficient resolution for the separation between the PLGA samples labeled to be 10.5 kDa and 21 kDa. To improve separation using this type of column, a column with a larger length could potentially be applied for the separation of the PLGA particles, to increase the plate number according to Eq. (2). To confirm that the low resolution is the cause of low column efficiency in the separation of these PLGA samples, a calibration curve could be constructed, which shows that an insufficient resolution is achieved. This particular calibration curve of the PLGel Mixed-E column was not made in this project.

-200 0 200 400 600 800 1,5 1,6 1,7 1,8 1,9 2 2,1 2,2 2,3 2,4 2,5 Si gna l (nRIU ) Time (min)

PLGel Mixed-E Column

(25)

25

Figure 11: Chromatogram of PLGA mixture, blank THF sample and toluene marker,

separated with a PLGel Mixed-D column.

Figure 11 shows the chromatogram of the same PLGA sample mixture, this time separated on

a PLGel Mixed-D column, which was tested due to its molar weight operating range of 200-400.000 g/mol and longer column length compared to the PLGel Mixed-E column, which in theory results in a higher plate number.19 It is assumed that the signal at 6.76 min. is corresponding to the PLGA mixture, due to its absence in the blank and because of elution before the toluene marker at 9.47 min. As can be observed from this figure, the PLGA sample mixture once again shows a broad distribution which is expected due to the polydispersity of PLGA polymers. Furthermore, it is assumed that the negative signals shown in the chromatogram are system peaks which is most likely the result of compounds in the THF mobile phase, which might be due to the formation of peroxides. Furthermore, from this figure it can be seen that the toluene marker has a significantly higher intensity of negative signals compared to the blank and PLGA samples. This can be related to the fact that the used vial with toluene, was prepared from a different bottle of THF solvent than the blank and PLGA sample. From this chromatogram it can be concluded that the composition of this THF was different. In Figure 12 below, a chromatogram is given which highlights the PLGA signal.

-5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 0 2 4 6 8 10 12 14 16 In ten sity (n RIU ) Time (min)

PLGel Mixed-D Column

Blank t0 PLGA

(26)

26

Figure 12: Chromatogram of PLGA mixture, separated with a PLGel Mixed-D column.

Using this chromatogram, it can be concluded that the PLGA mixture cannot be sufficiently separated with this column, due to the substantial broadness of the signals, resulting in low resolution. There is a small bump visible at 6.5 min which might correspond to the larger molecular weight PLGA sample, but there is a substantial overlap between both compounds. To increase separation, a column with higher column efficiency for this molecular weight range should be used.

separated with a PLGel Mixed double column.

-100 -50 0 50 100 150 200 250 300 5 5,5 6 6,5 7 7,5 8 In ten sity (n RIU ) Time (min)

PLGel Mixed-D Column

PLGA -4000 -2000 0 2000 4000 6000 8000 10000 0 5 10 15 20 25 In ten sity (n RIU ) Time (min)

PLGel Mixed 2 columns

Blank t0 PLGA

(27)

27

Figure 13 shows the chromatogram of the PLGA mixture, separated on two PLGel mixed

columns. As can be observed from the chromatogram, the signal at 14.5 min corresponds to the PLGA mixture, which is determined from the PLGA sample and the blank mixture. It is assumed that the positive signal at 18.8 min and the negative signals are system peaks, which might be the result of the formation of peroxides due to aging of the THF solvent. Figure 14 below highlights the PLGA signal.

Figure 14: Chromatogram of the PLGA mixture and individual injections, separated with a

PLGel Mixed Double column.

As illustrated in this chromatogram, there is a broad signal visible which is assigned to the 21 kDa PLGA sample, and a signal with less intensity at ~15.5 min. that overlaps with this broad peak of higher intensity. It is assumed this signal corresponds to the PLGA with a mass of 10.5 kDa, due to the overlap with the individual injection. The lower intensity of this smaller PLGA sample might be the result of either higher polydispersity, pipetting error or insufficient homogenization of both samples. Nevertheless, these results suggest that the PLGel Mixed double column can induce adequate separation of these PLGA particles.

-200 -100 0 100 200 300 400 12 13 14 15 16 17 In ten sity (n RIU ) Time (min)

PLGel Mixed 2 columns

PLGA 10.5 kDa PLGA 21 kDa PLGA Mixture

(28)

28 4.1.2 SEC calibration with polystyrene and polymethylmethacrylate

Figure 15: Chromatograms of PMMA with molar weight distributions of 2000 and 24830

g/mol.

Figure 15 depicts the chromatogram of PMMA polymer calibration standards with molar

weight distributions of 2000 g/mol, and 24830 g/mol used with a double column of PLGel Mixed (7.5x600 mm), recorded with RID detection. As portrayed in the chromatogram, there are signals at 14.3 and 16.2 min. which have been assigned to the PMMA standards. The positive signal at 18.0 min. and the negative signals are assumed to be system peaks related to degradation of THF to peroxides in the prepared samples and the mobile phase, and correspond to the blank sample which consisted of THF. Furthermore, as can be seen from this chromatogram, the signal corresponding to the smaller polymer standard elutes after the larger polymer standard, which is expected due to the fact that smaller molecules enter the pores of the packing more easily and are therefore more retained in SEC.

-3000 -2500 -2000 -1500 -1000 -500 0 500 1000 1500 2000 0 5 10 15 20 25 In ten sity (n RIU ) Time (min)

PLGel Mixed 2 Columns

PMMA 2000 g/mol PMMA 24830 g/mol

(29)

29

Figure 16: Chromatograms of PS with molar weight distributions of 1990 and 19880 g/mol.

Figure 16 depicts the chromatogram of PS calibration standards with molar weight

distributions of 1990 and 19880 g/mol, recorded with the same setup as the PMMA calibration. As is depicted in this figure, there are positive signals at 13.8 and 16.0 min. corresponding to the PS samples, which is determined from the subtraction of the blank sample. The positive signal at 17.9 min. and the negative signals are assumed to be system peaks, which occur due to aging of THF in the prepared samples and in the mobile phase.

-8000 -6000 -4000 -2000 0 2000 4000 0 5 10 15 20 25 In ten sity (n RIU ) Time (min)

PLGel Mixed 2 columns

PS 1990 g/mol PS 19880 g/mol

(30)

30

Figure 17: Calibration curves of PLGel Mixed double column, created with PMMA and PS

standards.

Figure 17 shows the calibration curves of both the PMMA calibration and the PS calibration.

As can be seen from this figure, both calibration curves show a linear trend after the permeation limit corresponding to the SEC region of the curve, up to a molar weight of about 107_g/mol and are quite comparable. This result was expected due to the fact that even though PMMA and PS have different hydrodynamic sizes, the retention mechanism is for both standards the same in SEC. From these calibration curves it can be concluded that both PS and PMMA can be used to determine PLGA molecular weights under these conditions.

Table 2: PLGA molecular weight results calculated according to PMMA and PS calibration

curve fitting given in Figure 17.

Calibration PLGA 21 kDa PLGA 10.5 kDa

PMMA 5758.52 g/mol 11005.14 g/mol

PS 5733.71 g/mol 10751.71 g/mol 0 1 2 3 4 5 6 7 8 0 5 10 15 20 Lo g MW VR(mL)

PMMA & PS calibration curves of PLGel Mixed 2

columns

PMMA Calibration PS Calibration

(31)

31 In Table 2, the results are given of the molecular weight calculations with the injected PLGA samples of 21 kDa and 10.5 kDa, using a third order fitting. As can be concluded from this table, the calculations for PLGA with a molar weight of 10.5 kDa are fairly accurate, while the calculation results for PLGA with a molar weight of 21 kDa show largely different numbers than what was expected.

There are multiple scenarios that could explain the results, the PLGA corresponding to the molar weight of 21 kDa could have decayed, which would have resulted in a lower molecular mass and is a feasible explanation due to the biodegradable nature of PLGA.20_The other scenario is that the PS and PMMA calibrations were not representative for the PLGA molecules due to the different ways that the molecules fold, although this case is less likely because of the fact that both calibrations give nearly the same results. Another potential issue could be that the PLGA of 21 kDa was labeled incorrectly, or that the PLGA sample corresponding to the 21 kDa molecular weight was deposited into the wrong vial, which could explain the higher viscosity of the PLGA which was labeled to be the 10.5 kDa PLGA (as higher molecular weight polymers generally have a higher viscosity than lower molecular weight polymers).5, 21

(32)

32

4.2 Separation of PLGA mixture with experimental core-shell columns

4.2.1 Column efficiency determination with PLGA

separated with the Phenomex core-shell B109 column.

Figure 18 depicts the chromatogram of the 10.5 kDa and 21 kDa PLGA mixture, as well as the

toluene and blank sample. As is shown in the chromatogram, there are broad signals visible at 6.03 and 6.76 min., which are assigned to the PLGA mixture due to the absence of these signals in the blank sample. The signal which is shown as the PLGA and blank sample at 7.14 min., as well as the negative peaks are assumed to be system peaks, caused by the aging of THF. The signal at 7.00 min. is assigned to the toluene marker, as it corresponds to the predicted permeation limit calculated by Eq. (4). The signal at 7.39 min. is expected to be either an impurity in the toluene sample or caused by aging of the sample, as the toluene sample was prepared from a different THF bottle than the blank and PLGA samples, and can therefore show different signals.

-60000 -40000 -20000 0 20000 40000 60000 0 2 4 6 8 10 12 14 16 In ten sity (n RIU ) Time (min)

Phenomenex core-shell B109

Blank t0 PLGA

(33)

33

Figure 19: Chromatogram of the PLGA mixture and individual PLGA injections, separated

with the Phenomenex core-shell B109 column.

Figure 19 shows the chromatogram of the PLGA mixture and individual PLGA injections, with

fixed y- and x-axis at the PLGA signals. There is a good separation visible between the 10.5 kDa and 21 kDa-labeled PLGA samples at 6.03 and 6.76 min. It has been determined that these signals indeed correspond to the used PLGA samples, by comparing them to the separate injections of both polymers. It can be observed that baseline separation had not been realized, but the Phenomenex core-shell B109 column achieves the highest resolution from all investigated columns, which can be related to the fact that the stationary phase consisted of small core-shell particles, resulting in lower chromatographic dispersion and thus higher resolution.8 Also the flowrate may have a role in the different selectivity, as a flowrate of 0.2 mL/min was used, in contrast to the 1.0 mL/min flowrate that was used for the measurement with the PLGel columns and which can potentially reduce the plate height according to Eq. (3), depending on the effective constants.

-100 400 900 1400 1900 2400 2900 3400 3900 5,4 5,6 5,8 6 6,2 6,4 6,6 6,8 7 In ten sity (n RIU ) Time (min)

Phenomenex core-shell B109

PLGA 10.5 kDa PLGA 21 kDa PLGA Mixture

(34)

34 4.2.2 SEC calibration with polymethylmethacrylate

Figure 20: Calibration curve of core-shell B109 column, created with PMMA standards.

Figure 20 shows the calibration curve of the Phenomenex core-shell B109 column with PMMA

standards using the HPLC setup with the RID. As can be seen from the calibration curve, there is a linear region between the permeation limit and the molar weight of 104.4 g/mol, and a non-linear region afterwards, which might indicate the exclusion limit in which HDC effects occur. This HDC region is clearly visible on the calibration curve for this column with the used standards due to the small particles (98 Å poresize) in the stationary phase. Using a third order fitting, the PLGA sample sizes were subsequently calculated, and given in Table 3 below.

Table 3: PLGA molar weight calculated according to the PMMA calibration curve fitting

given in Figure 20.

Calibration PLGA 21 kDa PLGA 10.5 kDa

PMMA 600.67 g/mol 5996.6 g/mol

y = -34,306x3_{+ 131,7x}2_{- 174,44x + 82,699} R² = 0,997 0 1 2 3 4 5 6 7 0,8 0,9 1 1,1 1,2 1,3 1,4 Lo g MW VR(mL)

PMMA calibration curve of Phenomenex core-shell B109

column

PMMA Calibration Poly. (PMMA Calibration)

(35)

35 As can be seen from this table, the calculated molar weight distributions are not in line with the expected values, as both the PLGA of size of 21 kDa and 10.5 kDa have significant lower calculated molar weights, even though the samples have molecular weights that fall within the linear region of the calibration curve. The results do show that the 10.5 kDa PLGA has a higher molecular mass compared to the 21 kDa PLGA, which is in agreement to the results given in

Table 2 and might be an indication that the 21 kDa PLGA had been substantially decayed. The

unexpected calculated values can also be the consequence of the fact that the PMMA molecules have a different way in which they fold compared to PLGA in this column with sub-3 µm particles. A proposed solution would be to use a different set of calibration standards, as this result shows that PMMA is not capable of providing a calibration curve that is accurate.

4.2.3 SEC Calibration with 3 coupled columns

Figure 21: Calibration curve of core-shell B111, B104 and B109 mixed columns, created

with PS standards.

Figure 21 depicts the calibration curve of the 3 core-shell B111, B104 and B109 mixed

columns. As can be seen from this figure, multiple areas in the curve can be distinguished.

y = -0,9051x3_{+ 9,6288x}2_{- 36,563x + 52,733} R² = 0,9896 0,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 2,4 2,6 2,8 3 3,2 3,4 3,6 3,8 4 Lo g MW V_R(mL)

PS calibration curve of Phenomenex core-shell B111, B104 and

B109 columns

PS Calibration with Slalom PS Calibration

(36)

36 There is a linear range between a molecular weight of 103.3 g/mol and 104.3 g/mol, after which an elevation of the calibration curve occurs that continues in a linear fashion between 105.5 g/mol and 106.1 g/mol. These different ranges are caused by the combining of columns, these stationary phases contain different particle and pore sizes, which has the same effect as a long mixed column. It is expected that the linear molecular weight ranges are separated using a SEC mechanism. Above the second molecular weight range, SC is clearly observed which is related to the stretching of the higher molecular weight polymers in the stationary phase.9_Furthermore, the data was fitted using a 3th_{order curve and is fairly precise when leaving out the SC region.} It can be concluded that this set of columns are efficient at separating low-molecular weight polymers up to a molecular weight of 106.1 g/mol, using a SEC mechanism, but might show inaccuracy at the elevation area.

4.2.4 Analysis of large polydispersity PS

Figure 22: Chromatogram of large polydispersity PS, blank THF sample and toluene marker,

separated on three Phenomenex core-shell columns (B111, B104 and B109).

Figure 22 depicts the chromatogram of a large polydispersity PS sample, blank and toluene

marker. -2 0 2 4 6 8 10 12 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 In ten sity (mA U ) Time (min)

Phenomenex core-shell B111, B104, B109

Blank Toluene PS Broad

(37)

37 As can be seen from this chromatogram, the PS sample shows a large distribution which is characteristic for large polydispersity polymers, and is absent from the blank sample which indicates that this is indeed the polystyrene sample. It was expected beforehand that due to the molecular weight distribution overlap with the elevation of the calibration curve at a molecular weight of 105.2_{g/mol, the peak would elute in a non-Gaussian shape. This result shows that} this is not the case, which was unexpected.

To induce this non-Gaussian peak shape, a larger polydispersity polystyrene might need to be used. Furthermore, it can be seen that during the polystyrene injection, residue of the toluene sample is injected, which can be solved by running a longer needle wash.

4.2.5 Concluding remarks regarding SEC analysis

From all the SEC measurements performed, it can be concluded that the proposed method with the respective double columns and core-shell columns are indeed capable of separating the used PLGA samples, but to be able to make a definite conclusion it is necessary to use undegraded PLGA building blocks to validate the SEC method and apply these building blocks to prepare nanoparticles with known compositions. The focus will now go to HDC analysis of polystyrene samples.

(38)

38

4.3 HDC Measurements

Figure 23: Chromatogram of 3-NBS acid separated on an Agilent PL-PSDA Cartridge Type 1

column.

To test the performance of the used HDC stationary phase, column tests have been performed with NBS acid, uracil and potassium dichromate. Figure 23 depicts the chromatogram of 3-nitrobenzenesulfonic acid (3-NBS acid) separated on an Agilent PL-PSDA Cartridge Type 1 HDC column, performed under the same conditions as the column test given by the manufacturer, shown in Appendix 1. As can be seen from this chromatogram, there is a broad peak visible with a retention time at 3.732 min.

Furthermore, substantial tailing could be observed of several minutes, which is unexpected when compared to the column performance test supplied by the manufacturer which shows a sharp peak with no tailing. According to the predicted value of V0 calculated with Eq. (4), from which the t0 marker retention time is deduced, it can be seen that the expected elution time is 7.2 min. (using a porosity of 0.42). This calculation is an approximation, as the practical elution time given by the manufacturer is 8.7 min. Nevertheless, it can be concluded that the column performance test results conducted in this study are not in line with the column test results from the manufacturer.

Figure 24 and Figure 25 show the injection with uracil and potassium dichromate on the same

column, which are general t0 markers in HDC.

-1 0 1 2 3 4 5 6 7 8 9 0 2 4 6 8 10 12 In ten sity (mA U ) Time (min)

Agilent PL-PSDA Cartridge Type 1

(39)

39

Figure 24: Chromatogram of uracil separated on an Agilent PL-PSDA Cartridge Type 1

column, with a flowrate of 1.6 mL/min and retention time at 3.77 min.

Figure 25: Chromatogram of potassium dichromate separated on an Agilent PL-PSDA

Cartridge Type 1 column, with a flowrate of 1.7 mL/min and retention time at 2.86 min.

As can be seen from both chromatograms of uracil and potassium dichromate injections on the HDC column, both signals show broad signals with substantial tailing. Uracil has a retention time of 3.77 min. at a flowrate of 1.6 mL/min, while potassium dichromate has a retention time of 2.86 min. at a flowrate of 1.7 mL/min, which are early elution times for t0 markers on this column, calculated under the respective conditions according to Eq. (4).

-50 0 50 100 150 200 250 300 350 400 450 0 2 4 6 8 10 12 14 16 In ten sity (mA U ) Time (min)

Agilent PL-PSDA Cartridge Type 1

Uracil -10 0 10 20 30 40 50 60 70 80 0 1 2 3 4 5 6 7 8 In ten sity (mA U ) Time (min)

Agilent PL-PSDA Cartridge Type 1

(40)

40

Figure 23-25 therefore suggest that the column stationary phase had been degraded. A possible

explanation could be that the stationary phase was damaged during column conditioning. It can be concluded that the available Agilent PL-PSDA Cartridge Type 1 column could not be used for the analysis of polystyrene analysis. This issue could be resolved by using a replacement of this HDC column.

(41)

41

5. Conclusion

With the analysis of PLGA on the different SEC columns investigated, insight was gained on the various column efficiencies. It can be concluded that some columns are indeed more effective for the separation of the PLGA mixture with labelled 10.5 kDa and 21 kDa molar weight distributions. Especially the PLGel Mixed double column and the Phenomenex core-shell column showed good separation results, where both signals in the mixture could be clearly distinguished. When the retention times of the PLGA samples were related to the calibration curves prepared with PS or PMMA standards, the PLGel Mixed double column showed accurate molecular weight results for the 10.5 kDa PLGA sample, while the molecular weight calculation performed with the Phenomenex core-shell column gave results that were inaccurate for both 21 kDa and 10.5 kDa PLGA samples. The latter column can therefore not be used for molecular weight determinations under the conditions that are applied in this study. The high efficiency of the Phenomenex core-shell column can furthermore be related to the smaller particles in the stationary phase which decreases the plate height and therefore increases the resolution.

Moreover, it can be concluded that the PLGA with a molecular weight of 21 kDa had been degraded, which was confirmed by SEC molecular weight determination measurements with the PLGel Mixed double column and the Phenomenex core-shell column, while it shows that the 10.5 kDa PLGA sample still had the original molecular weight distribution. In this study it is also shown that there is no significant difference between using PS or PMMA as calibration standards, as the calibration curves resulting from injections with either of the polymers were similar.

To validate the proposed second dimension SEC methods for the comprehensive analytical MAnIAC system, HDC was performed with the Agilent PL-PSDA Cartridge Type 1 column, but as shown during the column test with 3-NBS acid, uracil and potassium dichromate, the performance of this column was insufficient due to the extensive tailing of peaks and retention times that did not correspond to the specifications of the manufacturer. Further HDC measurements could therefore not be conducted with this column to validate the SEC methods which were developed in this study.

(42)

42

6. Future Prospects

A priority in this study was testing out the right columns and conditions in which polymer mixtures can effectively be separated by the SEC mechanism. These findings can subsequently be used in the optimization of the envisaged comprehensive analytical system for the analysis of polymers. It is expected that especially the application of Phenomenex core-shell columns are a good choice for this system, due to the high flowrates that can be utilized which facilitates faster separations in the second dimension. Furthermore, the Phenomenex core-shell column proved to give the highest resolution of all the tested columns, despite the smaller void volume of core-shell columns compared to columns with entirely porous particles. For future work, it would be viable to use the 3 coupled core-shell columns to inject the PLGA samples, and determine its molecular weight distributions with undegraded PLGA building blocks.

An important stage in the development of the second dimension is the validation of the method. It was envisaged to perform this validation with HDC, but due to issues with the stationary phase of the HDC column, this was not performed. In future studies, it is advised to continue the planned validation with HDC, to see if the method is indeed capable of separating polymer nanoparticles with known polymer lengths and compositions.

An extension of this work would be investigating the use of RPLC or IEC columns instead of the SEC separation methods studied in this research. Using RPLC or IEC columns, other characteristics of polymers like hydrophobicity and charge can be determined.

The next step in the development of the comprehensive analytical system would be combining both first and second dimension separation methods. To do this, HDC and SEC conditions need to be optimized for both methods to achieve the best possible performance.

(43)

43

References

[1] American Chemistry Council. The Basics: Polymer Definition and Properties

https://plastics.americanchemistry.com/plastics/The-Basics/ (accessed May 20, 2016)

[2] Understanding Nano. Nanoparticle Applications and Uses.

http://www.understandingnano.com/nanoparticles.html (accessed May 20, 2016).

[3] DSM Corporate Public Affairs. DSM position on nanotechnologies.

https://www.dsm.com/content/dam/dsm/cworld/en_US/documents/nanotechnology.pdf

(accessed May 20, 2016).

[4] DSM. DSM in Paint & Coatings. https://www.dsm.com/markets/paint/en_US/home.html

(accessed May 21, 2016).

[5] Edam, R. Comprehensive Characterization of Branched Polymers. Ph.D. Dissertation, University of Amsterdam, Amsterdam, 2013.

[6] Harris, D. C. Quantitative Chemical Analysis, 8th ed.; W. H. Freeman and Company: New York, 2010.

[7] The Linde Group. High-Performance Liquid Chromatography (HPLC).

http://www.lindeus.com/en/industries/laboratories/analysis_and_measurement/high-performance_liquid_chromatography/index.html (accessed Jun 9, 2016).

[8] Uliyanchenko, E. Size-exclusion chromatography – from high-performance to ultra-performance. Anal. Bioanal. Chem. 2014, 406, 6087–6094.

[9] Striegel, A. M.; Yau, W. W.; Kirkland, J. J.; Bly, D. D. Modern Size-Exclusion Liquid

Chromatography, 2nd ed.; Wiley: Hoboken, 2009.

[10] Striegel, A. M.; Brewer, A. K. Hydrodynamic Chromatography. Annu. Rev. Anal. Chem. 2012, 5, 15–34.

[11] Shimadzu. Size Exlusion Chromatography. http://www.shimadzu.com/an/hplc/support/ lib/lctalk/55/55intro.html (accessed May 29, 2016).

[12] Shodex. Calibration Curve for LF-804 (THF Eluent: PS) https://www.shodex.com/en/ dc/06/05/83.html (accessed May 29, 2016).

[13] Eclipse Business Media ltd. Separation Science, HPLC Solutions #101: Estimating Column Volume. http://www.sepscience.com/Techniques/LC/Articles/842-/HPLC-Solutions-101-Estimating-Column-Volume (accessed June 24, 2016).

[14] Popovici, S-T.; Kok, W. Th.; Schoenmakers, P. J. Band broadening in size-exclusion chromatography of polydisperse samples. J. Chromatogr. A. 2004, 1060, 237–252.

[15] Preti, R. Core-Shell Columns in High-Performance Liquid Chromatography: Food Analysis Applications. Int. J. Anal. Chem. 2016, 2016, 9 p.

(44)

44 [16] American Chemical Society. 55th Annual Report on Research 2010.

https://acswebcontent.acs.org/prfar/2010/AC7.html (accessed Jun 9, 2016).

[17] Uliyanchenko, E. V., Ultra-performance polymer separations. Ph.D. Dissertation, University of Amsterdam, Amsterdam, 2012.

[18] Eppert, J. E. HPLC Troubleshooting, 1st ed.; Separation Service Berlin: Berlin, 2003. [19] Agilent Technologies. PLGel Mixed column specifications. https://www.agilent.com/ en-us/products/gpc-sec/gpc-sec-columns/organic-gpc/plgel-mixed (accessed Jun 28, 2016). [20] Makadia, H. K.; Siegel, S. J. Poly Lactic-co-Glycolic Acid (PLGA) as Biodegradable Controlled Drug Delivery Carrier. Polymers 2011, 3, 1377–1397.

[21] Sigma Aldrich. Reference: Polymer Properties https://www.sigmaaldrich.com/content/ dam/sigma-aldrich/docs/Aldrich/General_Information/viscosity.pdf (accessed Jun 5, 2016).

(45)

45 Appendix 1: Column performance report of Agilent PL-PSDA Cartridge Type 1 column

Optimization and development of a second-dimension separation method for the analysis of polymer nanoparticles by comprehensive two-dimensional liquid chromatography

Bachelor Thesis Scheikunde