MSc Chemistry
Analytical Track
Master Thesis
Characterization and size separation of
nanoparticles using Hydrodynamic Chromatography
coupled with Inductively Coupled Plasma Mass
Spectrometry
by
Marina Boersma
11286776
March 2018
Number of Credits 42
September 2017- March 2018
Daily supervisor:
Examiner:
Dr. R.J.B Peters
Dr. W.T. Kok
RIKILT
A
CKNOWLEDGEMENT
After 7 months of master research, I have fulfilled my Master in Chemistry at the University of Amsterdam. The master research opportunity I had at RIKILT was a great chance of learning more about the subject. I would like to express my great appreciation to Dr R. J. B Peters, my supervisor, for his guidance, encouragement and constructive feedback during this research. I would also like to thank Anna Undas for her valuable and constructive feedback on my writing skills. Their advice helped me to understand and learn more about nanoparticles, hydrodynamic chromatography and inductively coupled plasma mass spectrometry. My grateful thanks are also extended to all the people from the laboratory who gave support and advice. Finally, I would like the thank all the interns for the support and the great time. Wageningen, Marina Boersma 28-02-2018
A
BSTRACT
Nanotechnology is a quick growing field with a wide range of different applications. Nanoparticles have many benefits like the creation of new substances with new characteristics. However, their behaviour in humans and animals differs from dissolved chemicals. This means that the potential risk is unclear and needs to be identified as quickly as possible. A variety of different techniques are available to determine the size of nanoparticles. For example, asymmetrical flow field flow fractionation is often used. A hydrodynamic chromatography (HDC) method coupled online with inductively coupled plasma mass spectrometry (ICP-MS) has been developed for inorganic nanoparticle analysis. The HDC method separates the particles based on their size and ICP-MS measures the mass of a single nanoparticle. When separating nanoparticles with HDC there are certain aspects that have to be considered. The low resolution of the HDC technique must be overcome and the stability of the nanoparticles need to be considered. The combination of these techniques provides information about the size of the nanoparticle and the agglomeration condition. The usefulness of the method was demonstrated for the study of the behaviour of nanoparticles during consumption. HDC-ICP-MS gave insight into the agglomeration state of gold nanoparticles within 20 minutes.
L
IST OF ABBREVIATIONS
AF4 Asymmetrical flow field flow fractionation Au-NP Gold nanoparticle CE Capillary electrophoresis CHDC Capillary hydrodynamic chromatography d-hydro Hydrodynamic chromatography diameter DLVO Derjaguin Landau Verwey Overbeek d-sp Single particle inductively coupled plasma mass spectrometry diameter FFF Field flow fractionation HDC Hydrodynamic chromatography HF5 Hallow fibre flow field flow fractionation HPLC High pressure liquid chromatography ICP-MS Inductive coupled plasma mass spectrometry i.d. Internal diameter MS Mass spectrometry NIST National Institute of Standards and Technology NP Nanoparticle NTA Nano tracking analysis PHDC Packed hydrodynamic chromatography PS-NP Polystyrene nanoparticle RP HPLC Reversed phase high pressure liquid chromatography RF Radio Frequency SDS Sodium dodecyl sulphate SEC Size exclusion chromatography spICP-MS Single particle inductively coupled plasma mass spectrometry SPR Surface Plasmon Resonance TE Transport efficiency TRA Time resolved analysis UV Ultraviolet
T
ABLE OF
C
ONTENTS
Acknowledgements ... 1 Abstract ... 2 List of abbreviations ... 3 1 Introduction ... 1 2 Background information ... 3 2.1 Nanoparticles ... 3 2.2 Hydrodynamic chromatography ... 4 2.2.1 Packed column hydrodynamic chromatography ... 5 2.2.2 Wide bore capillary hydrodynamic chromatography ... 6 2.3 Other separation techniques for nanoparticles. ... 7 2.4 Inductively Coupled Plasma Mass Spectrometry as detection technique ... 8 2.4.1 ICP-MS an inorganic ionization source ... 8 2.4.2 spICP-MS calculations ... 10 2.5 Nano tracking analysis ... 13 3 Materials and methods ... 14 3.1 Chemicals ... 14 3.2 Instrumentation ... 15 3.3 Sample preparations ... 15 3.4 Analytical procedure ... 16 3.5 Data processing HDC-spICP-MS ... 17 4 Results and Discussion ... 19 4.1 The separation of polystyrene nanoparticles with packed HDC with UV detection ... 19 4.1.1 Decreasing sodium dodecyl sulphate concentration ... 21 4.2 Separation of nanoparticles with wide bore capillary HDC ... 22 4.2.1 Influence of mobile phase on the separation ... 22
5 Conclusion ... 38 References ... 39 List of figures ... 42 List of tables ... 43 Appendix ... 44 A. Influence of mobile phase on the separation of polystyrene ... 44 B. HDC-spICP-MS calibration line results ... 46
1
I
NTRODUCTION
Nanoparticle technology is a steadily growing field in chemistry, because of the increasing use of nanoparticles in food, cosmetics and pharmaceutical industries. While the technology is expected to have many benefits, the potential risk of nanoparticles is unclear and needs to be identified as quickly as possible to ensure responsible use1, 2. The toxicity of the nanoparticle (NPs) is not just a function of the mass but depends also on size, shape, composition and agglomeration and this makes it difficult to describe3. In order to understand the impact, it is essential to have a robust and sensitive method to analyse the NPs. NPs can be distinguished into two groups, inorganic and organic. Inorganic NPs contain only non-organic components such as silica, gold and silver. The compounds most frequently used according to Gray et al. and Vance et al. are silver, carbon, titania, silica, zinc oxide and gold. Organic NPs are mostly composed of lipids, proteins and carbohydrates4, 5. The difficulty of the characterization of NPs is the often low concentration in complex matrices such as food, soil or surface water. Other obstacles that are faced during NPs analysis are the possible modification of the physical states (primary particle, aggregate or agglomerate) during analyses and sample preparation. The low concentration is several orders of magnitude below the sensitivity of many techniques, such as dynamic light scattering, differential centrifugal sedimentation, and field flow fractionation6 and it is therefore important to develop analytical methods that can directly monitor NPs at low concentrations. Single particle inductively coupled plasma mass spectrometry (spICP-MS) is a method that is able to determine particle size and particle number concentration at environmentally relevant concentrations4. The advantages of measuring with spICP-MS are the low detection limite (down to the ng/L range), the selectivity (element specific analysis), the fact that limited sample preparation is needed, and the rapid measurements (typically 1 minute). Furthermore, the size and particle concentration can be measured simultaneously. The downside, certainly from the view of a toxicologist, is that spICP-MS gives no information about the agglomeration state of NPs7,4, i.e. it cannot differentiate between primary particles, aggregates and agglomerates. Another method to characterize NPs is hydrodynamic chromatography (HDC), the separation mechanism relies only on the size and not on the coating or surface of the particle8,9. The HDC column is packed with non-porous beads, and in between the beads laminar flow occurs. The particles are separated by their velocity through the streamlines of the laminar flow which are faster near the centre of the parabolic or Poiseuille-like flow profile10. Larger particles will elute first because they cannot fully access the slow flow region near the wall11,12. By combining HDC with
The objective of this study is to obtain information on the separation and size determination of NPs with different HDC techniques and experiment with the coupling of HDC with spICP-MS. New data processing methods are required for handling the HDC and spICP-MS data in combination. The usefulness of the combination of HDC with spICP-MS will be shown by identifying the behaviour of the NPs during human consumption. An in vitro digestion model is used to simulate the mouth, stomach and intestine state with artificial juices13, 14.
2
B
ACKGROUND INFORMATION
2.1 N
ANOPARTICLESNanoparticles are extremely interesting due to their technological advantages. NPs have been used since the fourth century when the famous cup of Rome was made of dichroic glass creating a glittering effect with gold or copper NPs. Nowadays, different types of NPs such as amorphous, magnetic, and metal NPs are applied in food, cosmetics and the pharmaceuticals. NPs enables us to control and create substances with new characteristics but their behaviour differs from the behaviour of dissolved chemicals. Nanoparticles or ultrafine particles are defined as particles between 1 and 100 nm size with a surrounding interfacial layer5, 15. The small size of NPs lead to a more reactive behaviour. The environmental release of NPs during production and use impacts the ecosystem and ultimately the human health16. During the years, the major issue in nanoscience was the polydispersity of the NPs. Nevertheless, relatively monodisperse NPs could be made in some cases but the fact that no two NPs are the same could still be a problem. However, control on particle size and shape of NPs is achieved in many cases such as for gold and silver NPs. One of the application of NPs is based on the property of strongly enhancing the local electromagnetic near field and the effect of Surface Plasmon Resonance (SPR) which can lead to extremely strong light absorption and scattering17. The colloidal NPs could be sufficiently understood on aspects as size, shape and compositions, but the following factors still need to be examined: the surface layer, bonding of the stabilizers to the inorganic core, and the stability of the NP18. During this study gold NPs (Au-NPs) were stabilized with citrate molecules as shown in Figure 1. It is unclear how the citrate is absorbed on the surface. The stability of the NPs maybe of influence on the separation and on the ability to agglomerate. The Derjaguin Landau Verwey Overbeek (DLVO) theory is often used to describe agglomeration behaviour13. The theory states that interaction between particles results from the electrostatic repulsion so called double-layer interaction and van der Waals interactions. The electrostatic repulsion is the repulsion of charged particles. When two particles approach each other the electrostatic repulsion increases and keep the particles in dispersion. However, the van der Waals interaction increases as two particles get closer19. The charge of the surface depends on the pH of the solution due to protonation and deprotonation of the carboxyl groups of citrate. When the pH is below 5 suppression of dissociation of carboxylic groups occurs, this leads to the absence of electrostatic stabilization and agglomeration is observed in low (0.01 M) salt concentrations20.
2.2 H
YDRODYNAMIC CHROMATOGRAPHYHydrodynamic chromatography (HDC) separates particles according to size. The first HDC-type mechanism has been described by DiMarzio & Guttman21, multiple decades ago. HDC is a rapid technique and has a broad range of applications including NPs. For example, narrow capillary HDC has the potential of separation polymeric and bio-macromolecules on size. The advantages of a capillary column instead of using a packed column is the higher size resolution, the lower cost, and it is easy to apply22, 23, 24. As described by Striegel and Brewer, HDC is performed in an open tube (capillary, CHDC) or in a column packed with solid or nonporous inert particles (PHDC). The separation is based on the parabolic or Poiseuille-like flow that develops with laminar flow, in an open tube or in between the beads of the packed column. The inert beads in the PHDC column should minimize non-HDC interactions like van der Waals interactions between the column and the analytes. The non-HDC effects are especially common in aqueous solvents but could be minimized through the addition of salts or surfactants to the mobile phase10. In laminar flow the fastest streamlines are in the middle of the tube, and the slowest are near the walls25. The centre of the larger analytes cannot approach the walls of the tube as closely as the centre of smaller analytes. Therefore, the larger particles remain near the centre where they experience faster flow as shown in Figure 2. The smaller particles experience the faster streamlines but also the slower ones and travel through the tube with a slower average velocity than larger sized particles. The elution pattern is therefore from large to small particles. Figure 2, separation mechanism in HDC, (a) a two-component sample, (b) the larger analyte remains near the centre. The parabolic flow regime can be described with Reynolds numbers, Parabolic flow prevents in the range of 1 to 100 in Reynolds numbers and for an open tube should the Reynold number be less or equal to 2000. The equation of the Reynolds number for CHDC is described below: 𝑅𝑒 =2𝑅%𝜌𝑢 𝜂
a
b
where 𝜌 is the density of the medium, 𝜂 the viscosity of the medium, 𝑅% the radius of the capillary and 𝑢 the average flow velocity of the medium. For PHDC the Reynold number does not depend on the diameter of the column but on the diameter of the packing particles dp: 𝑅𝑒 =𝑑*𝜌𝑢 𝜂 The average velocity of particles can be described as: 𝑢* = 𝑢(1 + 2𝜆 − 𝜆0) The parameter λ is the ratio of the effective radius of the analyte to the radius of the capillary. From this equation the dimensionless residence time τ of the analyte can be described as: 𝜏 = (1 + 2𝜆 − 𝜆0)34 This equation is idealized in which particles are described as nonrotating, hard spheres. In practice one portion of the particle experiences a faster velocity than the opposite side of the particle this causes the particle to rotate. Particles may not be spherically shaped and may be permeable10. To account for these factors a modified quadratic term is added: 𝜏 = (1 + 2𝜆 − 𝐶𝜆0)34 2.2.1 Packed column hydrodynamic chromatography The packing of the column is inert and impermeable to NPs and therefore the separation must occur in the interstitial or void space of the column. According to Small et al.11 this is a minor effect and at least three effects should be considered; the hydrodynamic effect, the ionic strength effect and the
2.2.1.2 The ionic strength effect of the mobile phase Instead of being only dependent on the particle diameter, the separation as shown by Small et al., depended also on the ionic strength of the mobile phase11. The separation between particle size was found to increase with decreasing ionic strength. The electrical double layer of NPs can lead to electrostatic interactions between the surface of the packing and the NPs. The electrostatic repulsion determines the distance between the packing and the NPs. This determines the position of the NPs in the flow and influences the average velocity of the NPs. In low ionic strength the NPs will be repelled from the surface and forced into the faster moving stream lines11. 2.2.1.3 Van der Waals effects The van der Waals interactions between NPs and column packing becomes more important when the distance decreases. The NPs will stick more to the column packing, when the interaction is reversible the NPs will elute later. This occurs because the double layer of the particle diminishes, and the particle could approach closer to the packing. When the ionic strength of the eluent is high enough the NPs peak will disappear because the particle is bound to the packing material11. 2.2.2 Wide bore capillary hydrodynamic chromatography In 1984 Kelleher and Trumbore27 described a method for determining the molar weight of nanoparticles by injecting a sample in a capillary with an internal diameter of tenths of a millimetre, i.e. a wide bore capillary. For the determination of the molecular weight the peak shape was used rather than the retention time. Therefore this method could offer size determination without proper size separation. The chromatogram shown in Figure 3 consists of two peaks for the same size, and according to Kelleher and Trumbore27 the first sharp peak is the result of the domination of the NPs near the centre of the tube. The second rounded peak results primarily from molecules diffused near the wall of the tube28, 27. The first sharp peak (convection peak) is defined as h 1 and the second peak (diffusion peak) as h2. Since for a given flow rate and tube length the curve shifts from convection dominance to diffusion dominance as the molecular weight decreases27. So according to the Kelleher
and Trumbore article, the ratio R between h1 and h2 could be a potential parameter to correlate peak
shape to molecular weight.
Figure 3, chromatogram produced by injecting 100 µL of 30nm PS-NP This method was used to analyse the absorption of caffeine on Laponite and lanthanide on natural macromolecules29. -0.0004 -0.0002 0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0 2 4 6 8 10 12 14 16 18 20 Si gn al in A U Time in min 30 nm Inject h1 h2 1.2tb 2.0tb
2.3 O
THER SEPARATION TECHNIQUES FOR NANOPARTICLES.
Asymmetrical flow field-flow fractionation (AF4) is often used in combination with ICP-MS (but not spICP-MS) to determine the particle size and concentration. Ionic Ag and other metal ions can pass the membrane due to their small size and are therefore not detected when this technique is used. The resolution of AF4 is generally better than that of HDC. Another technique that can be used in combination with ICP-MS is capillary electrophoresis (CE). CE is able to separate ions from NPs. This technique is sensitive to the surface charge of the particle. This means that particles of the same size but different surface coating have different interactions and show particle separations. According to Zhou et al. a standard HPLC-ICP-MS method is able to separate NPs from ions up to 100 nm diameter, but the resolution was poor30. A different approach is to use a reversed phase column as presented by Helfrich et al. The pore size of the column was 300 Å and 1000 Å and the separation mechanism was found to be similar to size exclusion chromatography(SEC). In order to optimise the separation, the mobile phase contained a phosphate buffer and sodium dodecyl sulphate (SDS). The addition of SDS was necessary, otherwise the analytes could not elute from the column due to adsorption to the packing material. According to Sortebier et al. an RP HPLC column is a promising tool for further investigation of NPs and their corresponding ion. The separation mechanism here is a combination of SEC and interaction effects. The eluent can be chosen in such a way that one of the interactions is eliminated7.2.4 I
NDUCTIVELYC
OUPLEDP
LASMAM
ASSS
PECTROMETRY AS DETECTION TECHNIQUEIn general, inductively coupled plasma mass spectrometry (ICP-MS) is used to detect metals. Single particle inductively coupled plasma mass spectrometry (spICP-MS), a special mode of ICP-MS, is able to detect metallic NPs and to determine the particle size and particle number concentration at low concentrations4. 2.4.1 ICP-MS as inorganic ionization source ICP-MS is now widely used for inorganic characterization, quantitative and qualitative elemental analysis of high accuracy and high sensitivity. The injector consists of three concentric quartz tubes. The coil surrounding the largest tube is powered by an RF generator that produces typically a power of 1.5-2.5 kW at 27-40 MHz resulting in an ionised argon gas plasma31. A schematic overview of the ionization source is displayed in Figure 4. The plasma is initially made at atmospheric pressure and by an electrical tesla spark. The spark leads to the ionization of the argon gas. The ions present in the coil will interact with the high-frequency oscillating inductive field, this field is created by the RF current in the coil. The electrons produced by the ionization of argon will undergo the same events until the argon ionization process is stable. The temperature of the plasma is typically around 10000oK and requires thermal isolation from the outer quartz tubes. This is achieved by introducing a high-velocity flow of argon of about 10 L/min. The sample is introduced to the stable plasma as a vapour or aerosol of droplets by an argon flow rate of 1 L/min. The sample goes through the central tube of the injector where it rapidly vaporizes and atomizes. The resulting atoms will spend several milliseconds in a temperature region of 5000 to 10000 ˚K, and most elements ionize to a positive single charge. The ions are then introduced in the mass analyser, in this case a quadrupole mass analyser31.
Figure 4, schematic diagram of an inductively coupled plasma source. Metal based NPs produce a cloud of metal ions in the plasma torch as shown in Figure 5. When introduced in the mass spectrometer this cloud is detected as a single pulse and allows for the determination of the NP mass32. Figure 5, the ionization of a nanoparticle in the plasma torch. reproduced from RIKILT intern document.
2.4.2 spICP-MS calculations ICP-MS measures the total metal concentration. The metal ions are detected based on their mass to charge ratio and the intensity of the signal is correlated to the amount of metal. The intensities are then related to the calibration curve which is based on standards with known metal concentrations. During traditional ICP-MS, multiple intensity readings are integrated by using a specified interval time of 0.3-1 seconds, also known as dwell time33. In samples containing dissolved metals, the ions will be distributed homogenously which results in an average signal in time for a specific metal ion concentration as shown in Figure 6 (top). If the sample contains NPs the ions are no longer distributed homogenously but form clouds of ions as discussed before. Instead of a constant flow of metal ions entering the MS, clouds or clusters of ions are formed as shown in Figure 6. If a low dwell time is used the clusters of ions result in a spike in the intensity as shown in Figure 6 (bottom). Each pulse corresponds to an individual NP. Figure 6, on the top a spICP-MS chromatogram of dissolved analyte is shown resulting in a continuous flow and on the bottom spICP-MS chromatogram of NPs is shown. The NPs enter the MS in clusters of ions resulting in a spike in intensity. The image is adapted from RIKILT intern document. The assumption behind spICP-MS is that each pulse represents a single particle. In that case, the frequency of the pulses is directly related to the number based concentration of particles (the number of particles per volume) while the intensity of the pulse is related to the mass of the particle and through the density with the size of the particle. Often samples have to be diluted to observe individual particles.
2.4.2.1 Relating pulse frequency to particle number concentration Relating pulse frequency to particle number concentration; the particle number concentration, Np (particles/mL) is related to the frequency of the particle events, ƒ(Ip) (number of pulses/ms)33. Using the following equation: 𝑁*= ƒ(𝐼*) 𝑉𝜂< Where V (mL/ms) is the flow rate and 𝜂< is the transport efficiency. 2.4.2.2 Determining the transport efficiency The transport efficiency is defined as the ratio of the sample that reaches the plasma. In the spray chamber, aerosols are formed from the sample and only a small fraction of these aerosols reach the plasma. The number of NPs that reach the plasma depends on the number concentration of the NPs in the sample and the sample flow unto the ICP-MS. The transport efficiency typically ranges from 2 to 10% depending on the type of nebulizer that is used. The transport efficiency, also called the nebulization efficiency, is calculated as: 𝜂<=𝑁* 𝐶*× 1000 𝑉
Where 𝜂< is the transport efficiency, 𝑁* is the number of particles detected in the time scan (min-1),
𝐶* is the particle number concentration (L-1), and 𝑉 is the sample flow (mL/min). The transport efficiency can also be determined indirectly via the collection of waste stream of the spray chamber. The transport efficiency is determined by comparing the waste volume to the sample uptake volume, calculated as: 𝜂<= Δ𝑡𝑜𝑡𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 Δ 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡
2.4.2.3 Relating pulse height to particle size The pulse height can be related to particle size by the use of reference NPs. The reference NPs with pre-determined size should have the same elemental composition as the unknown sample. For NPs with no suitable reference material an alternative method is developed to calculate the particle height33. The first step described by Pace et al. is to create a calibration curve. With this calibration curve the signal intensity from the instrument is correlated to the concentration of the analyte entering the MS. The second step is to relate the concentration of the analyte to the total analyte mass that enters the plasma33. In the following equation, the relationship between the concentration of the analyte C (µg/mL) and W the mass observed per event (µg/event) is described: 𝑊 = 𝜂<𝑉𝑡L𝐶
Where ηn is the transport efficiency, V (mL/ms) is the flow rate, and td (ms/event) is the dwell time.
The intensity of each individual pulse Ip (counts/event), can be inserted into the transformed calibration curve to determine the mass of the corresponding particle mp 16 as shown in the following equation: 𝑚*= 𝐼*𝑡L 𝑅𝐹NOP× 𝑉𝜂< 60 × 𝑀* 𝑀S
In the equation 𝑚* is the particle mass (ng), 𝐼* the particle signal intensity (cps), 𝑅𝐹NOP
the ICP-MS
response for a standard (cps µg-1 L-1), 𝑡 L dwell time in (s), 𝑉 sample flow (mL/min), 𝑀* is the molar mass of the NP material and 𝑀S is the molar mass of the analyte measured. Assuming that the NP has a spherical geometry, the following equation is used to calculate the particle diameter 𝑑*(nm), 𝜌* is the particle density (g/mL): 𝑑*= 6𝑚* 𝜋𝜌*×10 U V
2.5 N
ANO TRACKING ANALYSISThe characterization of NPs becomes more and more important and relevant. Nanoparticle Tracking Analysis (NTA) allows sizing of the NPs34,35. The sample is injected into a sample holder which contains a cell with a prism, a laser beam enters the cell through a prism and the prism diverts it into a narrow beam with high intensity. The NPs present in the sample holder scatters the incoming laser light which is then visualised by the microscope. The camera which is mounted on the microscope makes a video of the visualized particle which move under Brownian motion. The Brownian motion is the random motion of particles in a liquid or gas. The software determines the average distance moved by each particle in the x and y plane. With this value, the particle diffusion coefficient is determined. The hydrodynamic diameter of the particle can be identified by using the Stokes Einstein equation if the sample temperature and viscosity were known36. 𝐷X = 𝑇𝐾[ 3𝜋𝜂𝑑 The Brownian motion occurs in three dimensions but only two dimensions are observed, however it is possible to determine the hydrodynamic diameter from measuring the mean squared displacement. These movements were tracked and measured with the NTA, Figure 7 shows a schematic overview of the NTA setup.
3
M
ATERIALS AND METHODS
3.1 C
HEMICALSGold nanoparticles (Au-NPs). Stabilized in citrate Au-NPs of several sizes (Sigma-Aldrich, nominal diameters: 30, 50, 100, and 200 nm) were used as size calibration standards. The 50nm standard was labelled by National Institute of Standards and Technology (NIST) as reference material and is provided with a Certificate of Investigating clarifying the particle size. Polystyrene nanoparticles (PS-NPs). The 3000 series Nanosphere polystyrene standards were obtained from ThermoFisher Scientific (Bremen, Germany). The particle diameters that were used: 203, 102, 46, 40, 30 and 20 nm. Chemicals to prepare synthetic digestion fluids. All chemicals were obtained from Merck, except for MgCl2 , glucuronic acid, lipase (pig), and α-amylase (Bacillus species) these were obtained from Sigma-Aldrich, glucosamine hydrochloride was obtained from Calbiochem, mucin (pig) was obtained from Carl Roth and sodium chloride and uric acid were obtained from VWR. All dilutions were prepared with MilliQ water as described in Table 1. Table 1, the composition of the juices used during the in vitro digestion model, the amounts were based on 1000 mL
Saliva pH 6.8
Gastric pH 1.3
Duodenal pH 8.1 Bile pH 8.2
Inorganic
896 mg KCl 824 mg KCl 7012 mg NaCl 5259 mg NaCl200 mg KSCN 2752 mg NaCl 3388 mg NaHCO3 5785 mg NaHCO3
1021 mg NaH2PO4*H2O 302 mg CaCl2 80 mg KH2PO4 376 mg KCl
570 mg Na2SO4 306 mg NaH2PO4*H2O 564 mg KCl 150 µL 37% HCl
298 mg NaCl 306 mg NH4Cl 50 mg MgCl2*6H2O 167.5 mg CaCl2
1694 mg NaHCO3 6.5 mL 37% HCl 180 µL 37% HCl
151 mg CaCl2
Organic
200mg urea 650 mg glucose 100 mg urea 250 mg urea290 mg amylase 20 mg glucuronic acid 1 g BSA 1.8 g BSA
15 mg uric acid 330 mg glucosamine
hydrochloride
9 g pancreatin 30 g bile
25 mg mucin 85 mg urea 1.5 g lipase MilliQ water
MilliQ water 1 g BSA MilliQ water
2.5 g pepsin
3 g mucin
MilliQ water
Mobile phase. The mobile phase consisted of 1mM sodium dodecyl sulphate (SDS) obtained from Fluka, and 10mM ammonium acetate obtained from Sigma-Aldrich.
3.2 I
NSTRUMENTATIONHydrodynamic chromatography (HDC). The HDC system consists of a hydrodynamic column and the standard UPLC system (A-30 Altus PerkinElmer, Buckinghamshire, United Kingdom) equipped with an Ultra Violet detector (Model N2971036). The PL-PSDA HDC type 1 column (length 800 mm and diameter of 7.5 mm) is used during the separation of the NPs. The column is packed with non-coated, nonporous silica spheres (Agilent Technologies, Wokingham UK), the particle size range that can be separated by this column is 5 nm to 300 nm. The system is driven by Empower3 software from Waters. In the case of capillary HDC, the HDC column is replaced by PEEK tubing with an internal diameter of 0.762mm and a length of 15m. The eluent is a 1mM solution of sodium n-dodecyl sulphate (SDS) with a flow rate of 1.0 mL/min with a sample injection volume of 100 µL. The HDC was coupled with spICP-MS, the size was also confirmed by NTA. Single particle ICP-MS (spICP-MS). The ICP-MS a quadrupole based Thermo Scientific X-series 2 (Waltham, MA, USA) ICP-MS operated in single particle mode, equipped with a standard nebulizer and a quartz impact bead spray chamber. The system was operated at a forward power of 1400W and the gas flows were set as plasma, 13 L/min; nebulizer 1.1 L/min; auxiliary, 0.7 L/min. The sample flow rate to the nebulizer was set at 1.5 mL/min using the integrated peristaltic pump. The Thermo PlasmaLab software in time resolved analysis (TRA) mode was used for data acquisition. The dwell time was set at 14 ms with an acquisition time of 900 s per measurement. Nano Tracking Analysis (NTA). the Nanosight LM20 is equipped with a laser module viewing unit (LM10) with a class 1 laser at 635nm (red laser), the maximum power of the laser is <50mW. The prism LM12 Optical flat (NTA4000) was used. The system was operated with a software dongle with version 3.1 build of the NTA software.
3.3 S
AMPLE PREPARATIONSPS-NPs analysed by HDC-UV. The PS-NPs standards are dissolved in MQ water with concentrations of 10 mg/mL. The vortex centrifuge was not used because agglomeration could occur so, the samples are shaken for 5 sec. Au-NPs analysed by HDC-UV. The gold standards are dissolved in MQ water with concentrations of 1 mg/mL. The vortex centrifuge should be used for 5 sec.
Agglomeration experiment. 100 µg/L 60 nm Au-NPs dispersed in MilliQ with pH of 2 achieved with the addition of HCl. Salt concentration can induce agglomeration therefore 60 nm Au-NPs were dispersed in MilliQ water with 2752 mg/L NaCl, 824 mg/L KCl and 302 mg/L CaCl2 and a pH of 2. Multiple samples were taken in time (t=0, t=2, t=20, t=40, and t=60 min) and measured with HDC-spICP-MS. Samples were diluted with MilliQ water to a concentration of 1 µg/L. In vitro digestion experiment. The in vitro digestion model is a simulation of the digestion process in the human body14. Artificial saliva, gastric juice, duodenal and bile juice are prepared by addition of relevant salts, enzymes and other characteristic compounds and set at relevant pH values as described in Table 1. Au-NPs analysed by NTA. 0.5 ppm 60 nm Au-NPs dispersed in MilliQ water, the vortex centrifuge should be used for 5 sec. The 60 nm Au-NPs dispersed in MilliQ water was used as a reference, for the in vitro digestion experiment measured with the NTA. The saliva, gastric, duodenal bile and agglomeration samples should have a concentration of 1 ppm, the samples were diluted with MilliQ water.
3.4 A
NALYTICAL PROCEDUREPacked bed HDC experiments (PHDC). The PS-NPs were first detected with a UV detector. After a baseline separation for the different PS sizes the method was applied to the Au-NPs. For the separation of the different Au-NP sizes, multiple aqueous mobile phases are used. The flow rate was set at 1 mL/min with a pressure of ~80 bar, and a concentration of minimal 1 mg/L NP concentration is required. Capillary HDC experiments (CHDC). The PS-NPs were separated using CHDC and detected with UV. After a baseline separation for the different PS sizes the method was applied to the Au-NPs. For the separation of the different sizes, multiple aqueous mobile phases are analysed. Multiple flow rates were used. The NTA experiments. The sample was injected into the sample holder and the location of the beam could be adjusted. When the camera gain was set to the maximum value and the camera level was increased it was possible to locate the laser beam. The gain is lowered again until the particles were no longer overexposed. Underexposure could be checked with the camera level settings. When the software displays the message “dark” the camera level should be increased until the message disappears. The samples would be analysed in triplicate and in between each measurement the plunger of the sample syringe is pressed till the particles were visibly moving. This should yield a more reliable result since NPs were unevenly dispersed. The NTA first records a video of the moving NPs which than processed using a number of threshold settings. In vitro digestion model. The digestion consists of three steps. First a saliva step with a short incubation at 37oC, after this the gastric juice was added and incubated for 2 hours at 37oC. The last
step consists of the addition of duodenal and bile juice followed by an incubation of 2 hours at 37oC.
The Au-NPs 60 nm NPs will be tested in two different digestion models; one with a pH of 2.5 in the stomach which simulates the original situation (empty stomach), and one with a pH of 5 which represents a fed state (full stomach).
Before the incubation, all digestive juices are heated to 37 ± 2oC and carried out head over head rotator at 37 ± 2oC to simulate the temperature and movement of the human body. After the addition of the gastric juice, the pH is checked and adjusted if necessary with 1M NaOH or 37% HCl. The digestion starts after 3 mL saliva is added to 475 µL 60nm gold NP (Au-NPs final concentration 100 µg/L) and was incubated for 5 min. Afterwards, 6 mL of gastric juice is added and the pH was adjusted to 5.0 ± 0.5 and the solution was incubated for 2 hours. Lastly, 6 mL duodenal and 3 mL bile juice was added the pH was adjusted to 6.5 ± 0.5 the mixture was incubated for 2 hours. A second experiment is performed with the pH set at 2.5 ± 0.5 in the stomach phase13. After 5 minutes of incubation, a subsample is taken from the mouth step and is diluted with MilliQ water to a concentration of 1 µg/L. During the stomach and intestine step, 4 subsamples were taken after 5, 20, 60 and 120 min of incubation. The samples were diluted with MilliQ to a final concentration of 1 µg/L. To reduce uncertainty the samples were analysed on size on the same day with the HDC-spICP-MS and NTA, a schematic overview is displayed in Figure 8. Figure 8, a schematic overview of the in vitro digestion model using two different pH for the stomach
3.5 D
ATA PROCESSINGHDC-
SPICP-MS
The data received from the HDC-spICP-MS experiments can be processed to determine an HDC- diameter and a spICP-MS diameter. Before that filtering is applied to remove noise from the spICP-MS analysis and memory effect from the HDC separation. Raw data filter. To make data meaningful a filter was used to remove the impurities from the real Duodenal + bile juiceSaliva Gastric juice
Rotate 5min 37˚C pH 6.5 Rotate 2hrs 37 ˚C pH 5/or 2 Rotate 2hrs 37˚C pH 8 Sample 60nm Au
Take sample after 5 min Take sample after 5, 20, 60 and 120 min
Take sample after 5, 20, 60 and 120 min Mo u th St om a c h Int e st ine
The intensity for 60 nm would be around 16000. The first filter removes the sum of 6 data points higher than 20000 and the second filter the sum 0f 60 data points lower than 24000 were removed. This approach was also applied to 100 and 200 nm Au-NPs. HDC-diameter. The HDC-diameter is based on the retention time. The following calculation was used to calculate the HDC diameter (d-hydro): 𝑑 − ℎ𝑦𝑑𝑟𝑜 = 𝑦 − 𝑏 𝑎 0 where y is the retention time in min, b is the intercept of the calibration line and a is the slope from the calibration line. The calibration line is prepared by determining the retention time of Au-NPs with known size spICP-MS diameter. The single particle ICP-MS diameter (d-sp) is based on the mass of the individual particle and density and could be calculated with the following equation: 𝑚* = 𝐼*𝑡L 𝑅𝐹bc<× 𝑉𝜂<∗ 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑓𝑟𝑎𝑐𝑡𝑖𝑜𝑛 60 ∗ 1000000
Where 𝑚* is the particle mass (ng), 𝐼*the particle signal intensity (cps), 𝑅𝐹NOP
the ICP-MS response
for a standard (cps µg
-1L
-1),
𝑡L dwell time (s), 𝑉 sample flow (mL/min), and 𝜂< the transport
efficiency. From the mass the diameter can be calculated as: 𝑑 − 𝑠𝑝 = (6× 𝑚* 1×104h 𝜋 /𝜌*)i.kkk×10000000
Where d-sp is expressed in nm and 𝜌* is the particle density in g/cm3. Finally, both diameters are
4
R
ESULTS AND
D
ISCUSSION
The scope of this project is the development of a method for size determination of NPs with HDC coupled online with ICP-MS to determine the presence of agglomerates. The first step is to develop a better separation of the NPs with the HDC column. The separation on HDC-column occurs due to a flow that is applied to a column packed with nonporous beads. The second step is the size determination and characterisation of Au-NPS with HDC-spICP-MS. The developed system is demonstrated through agglomeration of Au-NPs and to monitor the behaviour of nanoparticles during ingestion.
4.1 T
HE SEPARATION OF POLYSTYRENE NANOPARTICLES WITH PACKEDHDC
WITHUV
DETECTIONTo optimise the separation of NPs different eluents were used. The results showed that for the PS-NPs the separation with water as mobile phase was better than when other eluents were used. For detailed results see Appendix A. Nevertheless, water was not recommended by Agilent due to the stability of the column. Agilent uses ≤ 5% dodecan-1-ol ethoxylated and ≤1.5% SDS as eluents for the HDC-packed column. Figure 9 shows the results of measuring PS-NPs with water, after the column was stabilized with 5 mM SDS. The results showed baseline separation between 20 and 30 nm, and 30 and 40 nm. The larger NPs were almost baseline separated. So, to conclude this was a promising result. This experiment was measured 30 min after an SDS 5 mM experiment, which contributed in the conditioning of the column.
Figure 10, the separation of PS-NPs in MilliQ water without column conditioning with SDS When comparing the results in Figure 9 with the result in Figure 10, the results in Figure 9 were significantly better. Therefore, we conclude that the stability of the column and thus the separation were effected by measuring with MilliQ water only which means that the mobile phase influenced separation of the PS-NPs. Multiple articles mentioned an eluent consisting of 10 mM SDS in water and therefore, 10 mM SDS was used as mobile phase37, 2. The results showed no baseline separation between 20, 30 and 40 nm particle sizes and overlap with 46 nm particles indicating that particles smaller than 46 nm cannot be qualitative analysed with this eluent because the size would be overestimated to be ±46 nm diameter. Figure 11, results of measuring PS-NPs with 10mM SDS as mobile phase
From the result it was concluded that the separation with only MilliQ as eluents had the best quality when the column was stabilised with SDS. This means that there still is a small SDS concentration present in the column after stabilization with 5 mM SDS and this small amount of SDS influenced the separation of PS-NPs. 4.1.1 decreasing sodium dodecyl sulphate concentration Based on the obtained result a small concentration SDS was still present in the column when measured with MilliQ water. To reconstitute the elution time of 20 and 200 nm PS-NPs shown in Figure 9 the SDS concentration is gradually decreased till the same elution time was achieved. In Figure 12 the points represent the elution time (x-as) of 20 and 200 nm (size in the y-as) with different SDS concentrations. The points with the red circle represent the elution time of 20 and 200 nm measured with MilliQ water and a stabilized column. The following concentrations were used: 1.25, 0.63, 0.31, 0.16, 0.08, 0.04, 0.02, 0.01 and 0.005 mM. The results show that when the SDS concentration is lower the particles elute earlier. The 20 nm measured with 0.01 mM SDS elutes at the same time as the 20 nm measured according to Figure 9. The 200 nm particle when measured with 0.005 mM SDS elutes closer to the 200 nm particle when measured with MilliQ water. When the column is stabilized with 5 mM SDS and measured with MilliQ a gradient occurs in time. Figure 12, the elution time of 20nm and 200nm with different SDS concentration as mobile phase The same concentrations as described above were used in a gradient to separate PS-NPs. The
Figure 13, PS particle measured with a gradient (initial 100% 0.01mM SDS, 15min 100% 0.002mM SDS) In conclusion, PS-NPs in the range 20 to 203 nm could be baseline separated with the packed HDC column when a elution gradient was applied. An alternative method to separate NPs would be the use of a capillary column in HDC mode.
4.2 S
EPARATION OF NANOPARTICLES WITH WIDE BORE CAPILLARYHDC
The capillary HDC PEEK (polyether ether ketone) column from Agilent Technologies with the following dimensions 15 m x 0.03 inch I.D. (0.762 mm) was checked as alternative method. The method was based on the hydrodynamic flow profile and the diffusion coefficient of the particle. The diffusion coefficient depends on the diameter. Adsorption of the NPs to the column was less in capillary HDC than in packed HDC columns, therefore a technique with a smaller surface area could be helpful28. 4.2.1 Influence of mobile phase on the separation While the mobile phase should not influence the ability to separate different sized particles, the mobile phase may be of influence on the stability of the NPs and could therefore still influence the separation. The PS-NPs were eluted with 10 mM ammonium acetate with a flow rate of 0.8 mL/min as described in, Ya-Ru Tang et al.38. The results showed a small difference in peak shape and signal but not in elution time when compared with the results from PS-NPs measured in MilliQ with the same flow rate as shown in Figure 14.Figure 14, 30 nm PS-NP measured with MillQ water and 10mM ammonium acetate(AmAc) 4.2.2 Influence of the elution flow on the separation The sample is introduced in a laminar flow with a parabolic profile resulting in a peak with a sharp leading edge and a long decaying tail as shown in Figure 15. Nevertheless, when molecules have time to diffuse across the tube and thereby average their net velocity a Gaussian-shaped peak will be the result. This means that theoretically at a low flow rate more Gaussian-shaped like peaks would be likely.
between different NPs and was therefore out of the scope of this project. However the technique could be used for simple qualitative measurements29.
4.3 S
EPARATION OF GOLD NANOPARTICLES WITH PACKEDHDC
The results of PS-NPs as described in 4.1 showed that the mobile phase influenced the separation. To check if the same applies to Au-NPs, the Au-NPs were measured with MilliQ water and 10 mM SDS as mobile phase. The results in Figure 16 showed no baseline separation for Au-NPs when measured with MilliQ. Figure 16, measuring Au-NPs with MilliQ water as mobile phase The 10 mM SDS results did not give baseline separation either but the Au-NPs eluted around 5 minutes later compared to the Au-NPs measured with MilliQ water as eluent as show in Figure 17. To optimise the separation between different sizes, different SDS concentration were used as mobile phase. The Au-NPs were measured with 0.005, 0.01, 0.08 and 1 mM SDS concentration.Figure 17, Au-NPs separated with 10 mM SDS as mobile phase The Au particles measured with 0.005 mM SDS shown in Figure 18. The peaks were not separated and showed tailing. Figure 18, Au-NPs measured with 0.005 mM SDS as mobile phase Further, the NPs were measured with 0.01 mM and the improvement was observed when compared to the results with 0.005 mM SDS. This is an indication that the NPs were more stable in higher SDS concentrations. Next the concentration 0.08 mM SDS was tested and, results are shown in Figure 19. There was also an improvement in separation and less tailing occurred. The tailing depends on the diffusion of the particles in the faster and slower streamlines in the column, when diffusion is fast there will be less tailing.
Furthermore, a concentration of 1 mM SDS as mobile phase was analysed. The only difference between 0.08 mM SDS and 1 mM SDS as mobile phase is the elution time. The 1 mM SDS concentration is used for further analysis with the ICP-MS to achieve better column stability. In conclusion, the Au-NPs separation showed still some peak overlap. Nevertheless, a distinction could be made between lower than 30 nm and higher than 100 nm and even 50 nm could be partially identified, which is critical because the NP range applies to particles between 1 and 100 nm. Further analyses on size determination would be performed with 1 mM SDS eluent and a ICP-MS detector.
4.4 HDC-
SPICP-MS
For Au-NPs separation with HDC-spICP-MS the following method, a mobile phase of 1mM SDS and a flow of 1 mL/min was used. The ICP-MS detector is directly coupled with the HDC column. The ICP-MS is operated in single particle mode. With the following experiments two NP diameters will be determined, one according to the HDC retention time and one according to the spICP-MS data. Figure 20 shows the time scan that results from the analysis of 30 nm Au-NP. During this time we experienced some problems with the HDC analysis resulting in large memory peaks in the time scan. The 30 nm Au-NPs measured with HDC-spICP-MS should generate intensities up to ± 5000. However, the time scan in Figure 21 shows peak intensities up to 100000 cps resulting from unknown particles. Another reason to conclude that these higher peaks are some kind of noise is the fact that they appear randomly. Figure 20, raw data of 30 nm Au-NP measured with HDC-spICP-MS To make data meaningful a filter was used to remove the interferences from the real data resulting in the time scan shown in Figure 21. The HDC-UV results in paragraph 4.3 showed that the NPs elute from the column in a Gaussian shaped peak this means that single peaks could also be considered carry over and impurities. These peaks were excluded from the graph as shown in Figure 22. The Au-NPs could stick to plastic and glass and therefore to the tubing and column this gives carry over and impurities in the results.Figure 21, the 30 nm Au-NP after a filter is applied on the raw data A similar approach was used after the analysis of 60, 100 and 200 nm Au-NP. The time scans of these particles are given in Appendix B. The 30 nm Au-NP particles gives a peak intensity about 1500 cps. The 60 nm Au-NP is twice the size of the 30 nm Au-NP and thus 23 times the mass of the 30 nm Au-NP. Since the peak intensity correlates directly with particle mass we expect that the peak intensity for 60 nm Au-NP is 23 time 1500 cps is about 12000 cps. In the same way we expect a peak intensity of 3.33 times 1500 cps is about 55000 cps for the 100 nm Au-NPs. We find a peak intensity of around 60000 cps which confirms the expected peak intensity. At the same time we also notice the shift of retention time which is about 12.5 min for 30 nm Au-NP, 12.3 min for 60 nm Au-NP, 12.1 min for 100 nm Au-NP an about 11.9 min for 200 nm Au-NP as shown in Figure 22. The ionic gold showed a retention time of 12.8 min. Figure 22, Au-NPs measured with HDC-spICP-MS 0 20000 40000 60000 80000 100000 120000 140000 160000 180000 10 11 12 13 14 si gn al in c ps time in min 30 nm 60 nm 100 nm 200 nm
4.4.1 The calibration curve for gold nanoparticles The HDC column separates the NPs according to size and the order of elution is from large to small. This means that the size could be correlated with the retention time, therefore a calibration curve is made with known NPs. The relation between size of the NP and elution time is not linear and therefore a polynomial function is used as trend line as shown in Figure 23. Figure 23, calibration curve of Au-NP measured with HDC Further, the size was replaced by the size^0.5 to make the relation linear as shown in Figure 24, this formula was used to calculate de hydrodynamic diameter. y = 2E-05x2- 0,009x + 12,777 R² = 0,99988 11,8 11,9 12 12,1 12,2 12,3 12,4 12,5 12,6 12,7 12,8 12,9 0 50 100 150 200 250 tim e in m in size in nm y = -0,0629x + 12,805 R² = 0,98013 12,2 12,3 12,4 12,5 12,6 12,7 12,8 12,9 tim e in m in
30, 60, 100 and 200 nm Au-NP were analysed and the calculated hydrodynamic diameters (from HDC) and spherical equivalent diameters (from spICP-MS) were plotted in Excel resulting in the plot in Figure 25. Each particle size results in a cloud of points around a centre that is on a 45° angle which is exactly what we expect. The results also show that there is more spreading in the HDC direction then in the spICP-MS direction. Figure 25, calibration curve of Au-NPs containing both HDC-diameter (nm) and spICP-MS-diameter (nm). This method could be applied to analyse size. The experiments acquire two different size determination in one experiment which makes the results more reliable. Also if there is a large difference between hydrodynamic chromatography diameter (d-hydro) and the single particle ICP-MS diameter (d-sp) it could be an indication for agglomeration of the NPs.
4.5 HDC-
SPICP-MS
METHOD IN APPLICATION4.5.1 Agglomeration of Au-NPs The Au-NPs were able to agglomerate at low salt concentrations and at low pH. The spICP-MS method gives no information about the agglomeration state of NPs and it cannot differentiate between primary particles, aggregates and agglomerates. An aggregate consists of several primary particles bonded chemically while in an agglomerate the particles are physically bonded. An aggregate and agglomerate are therefore a number of particles with empty space between them. During the HDC analysis the aggregate is measured as a whole and the measured diameter will be the hydrodynamic diameter. In the spICP-MS we measure the spherical equivalent diameter, i.e. as if it is a solid particle. In the spICP-MS is the space between the primary particles not considered as shown in Figure 26. Therefore, in case of aggregates and agglomerates, the spherical equivalent diameter from the spICP-MS analysis is always smaller than the hydrodynamic diameter. To identify aggregates and agglomerates the HDC and spICP-MS are directly coupled. Figure 26, overview of aggregate and agglomerate diameter. Image adapted from Peters et al13. The results of such an analysis are shown in Figure 27. In this case 60 nm Au-NPs were allowed to agglomerate for 120 seconds at a pH of 1. Along the x-axis we see the hydrodynamic diameter of each particle (each dot is a measured particle) and along the y-axis we see the spherical equivalent diameter measured by spICP-MS. The cross hairs show that a HDC diameter of 300 nm corresponds to a spICP-MS diameter of 170 nm. The density (as m/V) of the agglomerate is an indication for the ‘density’ of the packing of the agglomerate, i.e. the ratio between total volume and the volume of solid material. The density relates to mass and volume and therefore to the third power of the DHDC DspICPMS
Figure 27, agglomeration of 60 nm Au-NPs at pH of 1 measured after 120 sec with HDC-spICP-MS The 60 nm Au-NPs is also exposed to low salt concentration in combination with low pH the result of which is shown Figure 28. The HDC-spICP-MS measurement showed that most particles were still in their primary state and some particles formed dimers. A small portion aggregated to a HDC diameter of 250 nm and a spICP-MS diameter of 150 which results in a packing ratio of 0.22 which is again an indication of loosely packed agglomerates. Figure 28, agglomeration of 60 nm at low salt concentrations and pH of 2 measured after 20 min with HDC-spICP-MS
The results of 60 nm Au-NP agglomerated in low concentration of salt and pH were also analysed with nano tracking analysis (NTA). The NTA results were shown in Figure 29 and conformed the assumption made based on Figure 28. Figure 29, 60 nm Au-NP at low salt concentration and a pH of 2 measured after 20 min with NTA 4.5.2 In vitro digestion of 60 nm Au-NPs One of the important questions in nano research into the presence of nanoparticles in food is what will happen with nanoparticles once they are ingested. Will they still be available for uptake of will they already dissolve in the stomach? Of course this also depends on the type of nanoparticle39. In this study we try to follow the dynamics of nanoparticles in the ingestion by exposing them to digestive juices in an in vitro digestion model. The in vitro digestion model is a simulation of the digestion process in the human body14, artificial saliva, gastric, duodenal and bile juice were used to study the behaviour of 60 nm Au-NP. The expectation was that the NPs would be stable to some extend during the saliva step and agglomerate during the gastric juice step because of the composition and the low pH value. Furthermore, during the duodenal and bile step it was expected that the agglomeration would reverse to its original state.
Figure 30, saliva sample taken after 5 min incubation at 37oC When the results were expressed in diameter the results in Figure 31 is obtained. Figure 31, saliva sample taken after 5 min expressed in hydrodynamic diameter and single particle diameter The results shown in Figure 31, appeared to have spreading in both directions. The diameter obtained with spICP-MS ranges from 50 to 100 nm diameter. The HDC results showed a diameter ranging from ± 30 to ± 150 nm. although we see some agglomeration it appears that most particles are still in their non-agglomerated state of 60 nm Au-NPs. This was also confirmed by measuring a saliva sample after 5 minutes with the NTA as shown in Figure 32.
Figure 32, NTA measurement of saliva sample with 1ppm NP concentration Subsequently, the gastric juice is added to the saliva juice and the pH is set at 2.0 ± 0.5. the HDC-spICP-MS results after 5 minutes and 120 minutes were displayed in Figure 33. Figure 33, gastric results after 5 min and after 120 min at a pH of 2.0 0 5000 10000 15000 20000 25000 30000 35000 0 2 4 6 8 10 12 14 Gastric, after 5 min 0 200 400 600 800 1000 1200 1400 1600 0 2 4 6 8 10 12 14 Gastric, after 120 min
Figure 34, 1 ppm gastric sample of pH 2.0 measured with NTA Following the gastric juice the intestine juices are added which bring the pH to a value of 8. During this phase 4 samples were collected after 5, 20, 60 and 120 min. Unfortunately due to a human error the sample of t=60 min was lost. The results of the three intestine samples are shown in Figure 35. After 5 min no Au-NPs are observed. However, after 20 min we already see the presence of peaks in the HDC-spICP-MS chromatogram. This was confirmed by even more particle peaks after 120 min. when combining the HDC and spICP-MS diameters in on graph as shown in Figure 36, it shows that there is a significant difference between the diameter obtained with the HDC and spICP-MS. While the hydrodynamic diameter ranges from ±50 to ±350 nm the spICP-MS diameter ranges from ±50 to ±150 nm. From this we conclude that the Au-NPS were still agglomerated and only a small part of the Au-NPS go back to their original 60 nm size. The packing ratio of the agglomerates is about 0.05 indicating very loosely packed agglomerates. It appears that the Au-NPs slowly transform from the large agglomerates in the gastric phase to smaller and loosely packed agglomerates in the intestine phase. This is of interest because may indicate that the Au-NPs will be available for uptake in the intestine.
5
C
ONCLUSION
The aim of this research project was to develop a method for the size separation and characterisation of nanoparticles (NPs) using hydrodynamic chromatography (HDC) coupled to the single particle inductively coupled plasma mass spectrometry (spICP-MS), overcoming the low resolution of HDC and the low concentrations of environmental samples. The HDC separation is based on the size of the NP and the type of column. However when polystyrene nanoparticles (PS-NPs) and gold nanoparticles (Au-NPs) of the same size were analysed with the same eluent the results were significantly different. The Au-NPs and PS-NPs stability was different and therefore two separate methods were developed. The best way to separate PS-NPs was with a gradient with SDS concentration decreasing form 0.01 to 0.002 mM SDS. The method used for Au-NPs was with a SDS concentration of 1 mM, the size range 200 to 10 nm could be analysed with this method. For higher size resolution a capillary HDC could be tested in future studies. The expected advantages of capillary HDC is the less eddy diffusion resulting in less peak broadening and more cost efficient. The HDC method for the separation of Au-NPs described above, was used for the HDC-spICP-MS experiments. An online method was developed combining packed HDC, to obtain particle size by means of elution time, and spICP-MS, to obtain particle size by the mass of the particle. When Au-NPs were allowed to agglomerate, the state of the agglomeration could be identified by the packing ratio and the NPs were characterised. The behaviour of Au-NPs during consumption is analysed with HDC-spICP-MS by the use of an in vitro digestion model. Three stages are simulated: mouth, stomach and intestine. The Au-NPs form large agglomerates in the stomach phase. During the intestine phase the Au-NPs transform from the large agglomerates in the gastric phase to smaller and loosely packed agglomerates. This is of interest because may indicate that the Au-NPs will be available for uptake in the intestine. It can be concluded that the coupling of HDC and spICP-MS proved sufficient separation on size and is able to characterize Au-NPs.
R
EFERENCES
(1) Tiede, K.; Boxall, A. B. A.; Wang, X.; Gore, D.; Tiede, D.; Baxter, M.; David, H.; Tear, S. P.; Lewis, J. Application of Hydrodynamic Chromatography-ICP-MS to Investigate the Fate of Silver Nanoparticles in Activated Sludge. J. Anal. At. Spectrom. 2010, 25, 1149. (2) Pergantis, S. A.; Jones-Lepp, T. L.; Heithmar, E. M. Hydrodynamic Chromatography Online with Single Particle-Inductively Coupled Plasma Mass Spectrometry for Ultratrace Detection of Metal-Containing Nanoparticles. Anal. Chem. 2012, 84, 6454–6462. (3) López-Serrano, A.; Olivas, R. M.; Landaluze, J. S.; Cámara, C. Nanoparticles: A Global Vision. Characterization, Separation, and Quantification Methods. Potential Environmental and Health Impact. Anal. Methods 2014, 6, 38–56. (4) Gray, E. P.; Bruton, T. A.; Higgins, C. P.; Halden, R. U.; Westerhoff, P.; Ranville, J. F. Analysis of Gold Nanoparticle Mixtures: A Comparison of Hydrodynamic Chromatography (HDC) and Asymmetrical Flow Field-Flow Fractionation (AF4) Coupled to ICP-MS. J. Anal. At. Spectrom. 2012, 27, 1532. (5) Vance, M. E.; Kuiken, T.; Vejerano, E. P.; McGinnis, S. P.; Hochella, M. F.; Hull, D. R. Nanotechnology in the Real World: Redeveloping the Nanomaterial Consumer Products Inventory. Beilstein J. Nanotechnol. 2015, 6, 1769–1780. (6) Rakcheev, D.; Philippe, A.; Schaumann, G. E. Hydrodynamic Chromatography Coupled with Single Particle-Inductively Coupled Plasma Mass Spectrometry for Investigating Nanoparticles Agglomerates. Anal. Chem. 2013, 85, 10643–10647. (7) Sötebier, C. A.; Weidner, S. M.; Jakubowski, N.; Panne, U.; Bettmer, J. Separation and Quantification of Silver Nanoparticles and Silver Ions Using Reversed Phase High Performance Liquid Chromatography Coupled to Inductively Coupled Plasma Mass Spectrometry in Combination with Isotope Dilution Analysis. J. Chromatogr. A 2016, 1468, 102–108. (8) Tiede, K.; Boxall, A. B. A.; Tiede, D.; Tear, S. P.; David, H.; Lewis, J. A Robust Size-Characterisation Methodology for Studying Nanoparticle Behaviour in “real” Environmental Samples, Using Hydrodynamic Chromatography Coupled to ICP-MS. J. Anal. At. Spectrom. 2009, 24, 964. (9) Philippe, A.; Schaumann, G. E. Evaluation of Hydrodynamic Chromatography Coupled with Uv-Visible, Fluorescence and Inductively Coupled Plasma Mass Spectrometry Detectors for Sizing and Quantifying Colloids in Environmental Media. PLoS One 2014, 9, 1–9. (10) Small, H.; Langhorst, M. A. Hydrodynamic Chromatography. Anal. Chem. 1982, 54, 892A– 898A.
(14) Bove, P.; Malvindi, M. A.; Kote, S. S.; Bertorelli, R.; Summa, M.; Sabella, S. Dissolution Test for Risk Assessment of Nanoparticles: A Pilot Study. Nanoscale 2017, 9, 6315–6326. (15) Batista, C. A. S.; Larson, R. G.; Kotov, N. A. Nonadditivity of Nanoparticle Interactions. Science (80-. ). 2015, 350. (16) Peters, R. J. B.; van Bemmel, G.; Milani, N. B. L.; den Hertog, G. C. T.; Undas, A. K.; van der Lee, M.; Bouwmeester, H. Detection of Nanoparticles in Dutch Surface Waters. Sci. Total Environ. 2018, 621, 210–218. (17) Fratoddi, I.; Cartoni, A.; Venditti, I.; Catone, D.; O’Keeffe, P.; Paladini, A.; Toschi, F.; Turchini, S.; Sciubba, F.; Testa, G.; et al. Gold Nanoparticles Functionalized by Rhodamine B Isothiocyanate: A New Tool to Control Plasmonic Effects. J. Colloid Interface Sci. 2018, 513, 10–19. (18) Jin, R.; Zeng, C.; Zhou, M.; Chen, Y. Atomically Precise Colloidal Metal Nanoclusters and Nanoparticles: Fundamentals and Opportunities. Chem. Rev. 2016, 116, 10346–10413. (19) Verwey, E. J. W.; Overbeek, J. T. G. Theory of the Stability of Lyophobic Colloids. J. Colloid Sci. 1955, 10, 224–225. (20) Horvolgyi, Z.; Kiss, E. Progress in Colloid and Polymer Science; 135th ed.; Springer: Germany, 2008. (21) Flow, S. B. Y. POLYMER L E T T E R S VOL. 7 , PP. 267-272 (1969) SEPARATION BY FLOW In T H I S Note We Propose. 1969, 7, 267–272. (22) DosRamos, J. G.; Silebi, C. A. The Determination of Particle Size Distribution of Submicrometer Particles by Capillary Hydrodynamic Fractionation (CHDF). J. Colloid Interface Sci. 1990, 135, 165–177. (23) Brough, A. W. J.; Hillman, D. E.; Perry, R. W. Capillary Hydrodynamic Chromatography --an Investigation into Operational Characteristics. J. Chromatogr. A 1981, 208, 175–182. (24) Liu, L.; Veerappan, V.; Pu, Q.; Cheng, C.; Wang, X.; Lu, L.; Allen, R. D.; Guo, G. High-Resolution Hydrodynamic Chromatographic Separation of Large DNA Using Narrow, Bare Open Capillaries: A Rapid and Economical Alternative Technology to Pulsed-Field Gel Electrophoresis? Anal. Chem. 2014, 86, 729–736. (25) Blom, M. T.; Chmela, E.; Gardeniers, J. G. E.; Tijssen, R.; Elwenspoek, M.; Van den Berg, A. Design and Fabrication of a Hydrodynamic Chromatography Chip. Sensors Actuators, B Chem. 2002, 82, 111–116. (26) DiMarzio, E. A.; Guttman, C. M. Separation by Flow. Macromolecules 1970, 3, 131–146. (27) Trumbore, F. M. K. Dispersion of Protein Solutions in Short , Open Capillary Tubes as a Method for Rapid Molecular Weight Determination Peak Shapes as a Function Qf Flotil Rate. 1984, 24, 20–24. (28) Fischer, C. H.; Giersig, M. Analysis of Colloids. VII. Wide-Bore Hydrodynamic Chromatography, a Simple Method for the Determination of Particle Size in the Nanometer Size Regime. J. Chromatogr. A 1994, 688, 97–105. (29) Sorauf, K. J.; Connors, D. E.; Wells, T. A.; Miller, K. E. A Hydrodynamic Method for the Measurement of Laponite-RD Caffeine Binding. Appl. Clay Sci. 2014, 87, 197–204.
(30) Zhou, X. X.; Liu, R.; Liu, J. F. Rapid Chromatographic Separation of Dissoluble Ag(I) and Silver-Containing Nanoparticles of 1-100 Nanometer in Antibacterial Products and Environmental Waters. Environ. Sci. Technol. 2014, 48, 14516–14524. (31) Davies, A. N. Mass Spectrometry; 2005. (32) Peters, R.; Herrera-Rivera, Z.; Undas, A.; van der Lee, M.; Marvin, H.; Bouwmeester, H.; Weigel, S. Single Particle ICP-MS Combined with a Data Evaluation Tool as a Routine Technique for the Analysis of Nanoparticles in Complex Matrices. J. Anal. At. Spectrom. 2015, 30, 1274–1285. (33) Pace, H. E.; Rogers, N. J.; Jarolimek, C.; Coleman, V. A.; Higgins, C. P.; Ranville, J. F. Determining Transport Efficiency for the Purpose of Counting and Sizing Nanoparticles via Single Particle Inductively Coupled Plasma Mass Spectrometry. Anal. Chem. 2011, 83, 9361– 9369. (34) Lawler, D. F.; Youn, S.; Zhu, T.; Kim, I.; Lau, B. L. T. Comprehensive Understanding of Nano-Sized Particle Separation Processes Using Nanoparticle Tracking Analysis. Water Sci. Technol. 2015, 72, 2318–2324. (35) Worldwide, M. I. WHITEPAPER NTA : Principles and Methodology. 2015, 1–16. (36) Hole, P.; Sillence, K.; Hannell, C.; Maguire, C. M.; Roesslein, M.; Suarez, G.; Capracotta, S.; Magdolenova, Z.; Horev-Azaria, L.; Dybowska, A.; et al. Interlaboratory Comparison of Size Measurements on Nanoparticles Using Nanoparticle Tracking Analysis (NTA). J. Nanoparticle Res. 2013, 15. (37) Helsper, J. P. F. G.; Peters, R. J. B.; Brouwer, L.; Weigel, S. Characterisation and Quantification of Liposome-Type Nanoparticles in a Beverage Matrix Using Hydrodynamic Chromatography and MALDI-TOF Mass Spectrometry. Anal. Bioanal. Chem. 2013, 405, 1181–1189. (38) Tang, Y. R.; Huang, H. Y.; Hu, J. B.; Rattinam, R.; Li, C. H.; Chen, Y. C.; Urban, P. L. Capillary Hydrodynamic Chromatography Reveals Temporal Profiles of Cell Aggregates. Anal. Chim. Acta 2016, 910, 75–83. (39) Braakhuis, H. M.; Kloet, S. K.; Kezic, S.; Kuper, F.; Park, M. V. D. Z.; Bellmann, S.; van der Zande, M.; Le Gac, S.; Krystek, P.; Peters, R. J. B.; et al. Progress and Future of in Vitro Models to Study Translocation of Nanoparticles. Arch. Toxicol. 2015, 89, 1469–1495. (40) Toyohide TAKEUCHI, SISWOYO, Zarina ASPANUT, L. W. L. Hydrodynamic Chromatography of Macromolecules on Small Spherical Non-Porous Silica Particles. J. Chromatogr. A 506, 547– 561. (41) Product Information; Saint Louis, MO, 2016.
