Dynamics of individual magnetic particles near a biosensor
surface
Citation for published version (APA):
van Ommering, K. (2010). Dynamics of individual magnetic particles near a biosensor surface. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR658494
DOI:
10.6100/IR658494
Document status and date: Published: 01/01/2010 Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
The work described in this thesis has been carried out at the Philips Research Laboratories Eindhoven, The Netherlands, as part of the Philips Research program.
Copyright © 2010 by K. van Ommering
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission from the copyright owner. A catalogue record is available from the Eindhoven University of Technology Library. ISBN: 978-90-386-2179-1
Dynamics of Individual Magnetic Particles
near a Biosensor Surface
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus, prof.dr.ir. C.J. van Duijn, voor een
commissie aangewezen door het College voor Promoties in het openbaar te verdedigen
op maandag 14 juni 2010 om 16.00 uur
door
Kim van Ommering geboren te Geldrop
Dit proefschrift is goedgekeurd door de promotor: prof.dr.ir. M.W.J. Prins
Copromotor:
Om Asato Ma Sad Gamaya
(“Lead us from Ignorance to Truth”)
Table of Contents
Summary iii Samenvatting vii Preface xi Chapter 1 Introduction 1 1.1 Biomolecular Diagnostics 1 1.2 Point-of-Care Biosensors 2 1.3 Magnetic Particles 41.4 Detection of Magnetic Particles 6
1.5 Research Goal 10
1.6 Magnetic Particle Characterization 12
1.7 Confined Brownian Motion Analysis 15
1.8 Bond Characterization 17
1.9 Out-of-Plane Mobility in Microscopy 19
1.10 Thesis Outline 21
Chapter 2 Magnetic Particles 29
2.1 Magnetism 29
2.2 Ferrimagnetism, Domains and Anisotropy 30
2.3 Superparamagnetism 32
2.4 Superparamagnetic Bead 33
2.5 List of Used Particles 36
Chapter 3 Confined Brownian Motion Analysis 39
3.1 Introduction 40
3.2 Materials and Methods 40
3.3 Results 42
3.4 Conclusions 44
Chapter 4 Individual Magnetic Particle Motion 47
4.1 Introduction 48
4.2 Materials and Methods 48
4.3 Magnetophoretic Analysis 49
4.3.1 Crossing between two wires 50
4.3.2 High frequency fields 53
4.3.3 Wire channel set-up 54
4.4 Confined Brownian Motion Analysis 56
4.4.1 Wire with surface barriers 56
4.4.2 Shaped wire 59
Chapter 5 Mobility of Bound Particles 69
5.1 Introduction 70
5.2 Theoretical Estimations 70
5.3 Materials and Methods 72
5.3.1 Detection 72
5.3.2 Model assay 74
5.4 Results and Discussion 75
5.4.1 Characterization of the set-up using fixated particles 75
5.4.2 Intensity fluctuations of bound particles 77
5.5 Conclusions 81
Chapter 6 Bond Characterization 85
6.1 Introduction 86
6.2 Materials and Methods 87
6.3 Geometric Description 89
6.4 Results 90
6.4.1 Particle mobility and bond flexibility 90
6.4.2 Mobility differences between particles 93
6.4.3 Influence of magnetic forces on particle mobility 96
6.4.4 Influence of magnetic anisotropy on particle mobility 98
6.5 Conclusions 103
Chapter 7 Electrostatic Height Modulation 109
7.1 Introduction 110
7.2 Materials and Methods 110
7.3 Results and Discussion 112
7.4 Conclusions 116
Curriculum Vitae 121
List of Publications 122
Dynamics of Individual Magnetic Particles
near a Biosensor Surface
S
ummary
The use of magnetic particles in biosensing is advantageous for transport of target molecules in a device, for assay integration, and for labeled detection. The particles generally have a size between 100 nm and 3 μm and are of a superparamagnetic nature, being composed of thousands of iron oxide grains in a polymer matrix. In this thesis we describe a series of detailed microscopy studies of magnetic particles near to and coupled to a biosensor surface, in order to characterize their dynamic behavior and their magnetic properties.
The first part of the thesis deals with the use of integrated microscopic current wires to study and manipulate unbound particles on a chip surface. The magnetic properties of individual particles are characterized in magnetic fields below 10 mT, using on-chip magnetophoretic analysis and on-chip Brownian motion analysis. In magnetophoretic analysis, the volume suscepti-bility of 1 µm particles is determined by optically measuring the speed of particles moving between two current wires. The analysis reveals distinct differences in volume susceptibilities of particles with the same outer diame-ter. In addition to DC magnetic fields, also AC magnetic fields are applied, showing a decrease in particle susceptibility for increasing field frequencies. To reduce the hydrodynamic perturbation by the surface in on-chip magneto-phoretic analysis, we present a chip design in which a particle can move back and forth in the channel between two large wires.
In Brownian motion analysis, small particles of 150 to 450 nm are trapped in a tunable magnetic potential well above an integrated current wire. The histogram of two-dimensional particle positions reveals the strength of the particle magnetization. Using straight current wires, we demonstrate differ-ences in bead susceptibility of an order of magnitude and differdiffer-ences in vol-ume susceptibility of more than a factor of two. By using wires with surface barriers and wires with a tapered shape, the measurement error in the suscep-tibility determination is decreased to less than 10%. We also show that com-bining a tapered wire with an external uniform field can give additional infor-mation on particle properties such as anisotropy or a small permanent mag-netic moment.
The second part of the thesis describes a study of the dynamics of particles that are biologically bound to a sensor surface. We show that an optical eva-nescent field can be used to study the thermal out-of-plane motion of bound particles with nanometer resolution, because the scattered light intensity of the particles depends on the height above the surface. By using a biological tether of known length (dsDNA of 290 bp/99 nm) we show that height varia-tions can be quantitatively determined. We demonstrate that the accuracy of the height determination depends on the properties of the used particles, e.g. the shape, smoothness, and internal structure. Optimal results are found for non-magnetic polystyrene particles and magnetic particles that are smooth and spherical.
Next, we show that the bond between a particle and a surface can be characterized by measuring the three-dimensional thermal mobility of the particle. As a model analyte we use four different lengths of dsDNA to bind the particle to the surface (590 bp/201 nm, 290 bp/99 nm, 141 bp/48 nm and 105 bp/36 nm). Plots of the minimum height, average height and maximum height as a function of the in-plane particle position reflect the differences in bond length, bond flexibility and bond orientation of the different DNA molecules. We also analyze ensembles of particles bound to the four DNA lengths and show that the height displacement is at maximum equal to the bond length, but large variations between particles are observed, which we attribute to non-specific interactions.
The mobility of a bound particle can be influenced by applied forces. We show that a magnetic gradient force towards the surface brings bound parti-cles on average closer to the surface. However, a magnetic gradient force away from the surface does not always brings bound particles away from the sur-face, but can lead to medium or minimum heights. This can be explained by magnetic anisotropy in the particles, leading to particle alignment and subse-quent height reduction. We describe a model that shows how particle align-ment due to magnetic anisotropy brings bound particles into well-defined three-dimensional positions. Although particle alignment may interfere with
the desired magnetic manipulation, it can also lead to additional information on for example rotational freedom of the bond and bond flexibility.
Finally, we show that a bound particle can also be pulled away from the surface by an electrostatic force induced by replacing the buffer with a buffer of low ionic strength. We show that the height modulation is dependent on both the analyte length and the ionic strength, and describe a quantitative model to account for the measured height displacements for particles bound to four different lengths of DNA.
The improved knowledge on the magnetic properties of individual parti-cles and the mobility of individual partiparti-cles bound to a biosensor surface, resulting from the experiments described in this thesis, may lead to improved detection limits and enhanced specificity in future biosensors.
Dynamica van Individuele Magnetische
Deeltjes dichtbij een Biosensor Oppervlak
S
amenvatting
Het gebruik van magnetische deeltjes in biosensoren is gunstig voor transport van moleculen, voor assay integratie en voor gelabelde detectie. De deeltjes hebben meestal een grootte tussen 100 nm en 3 μm en zijn superparamagne-tisch, omdat ze bestaan uit duizenden ijzeroxide korreltjes in een polystyrene matrix. In dit proefschrift beschrijven we een serie van gedetailleerde micro-scopie-experimenten met magnetische deeltjes dichtbij of gebonden aan een biosensoroppervlak. Hiermee karakteriseren we het dynamische gedrag van de deeltjes en hun magnetische eigenschappen.
Het eerste deel van dit proefschrift gaat over het gebruik van geïntegreer-de microscopische stroomdrageïntegreer-den, waarmee het gedrag van ongebongeïntegreer-den deeltjes dichtbij een chipoppervlak kan worden bestudeerd en gemanipuleerd. We karakteriseren de magnetische eigenschappen van individuele deeltjes in magnetische velden lager dan 10 mT met zowel magnetoforetische analyse als Brownse-bewegingsanalyse. Met magnetoforetische analyse bepalen we de volumesusceptibiliteit van 1 μm deeltjes door optisch de snelheid van de deeltjes te meten wanneer deze tussen twee stroomdraden heen en weer bewegen. De metingen laten aanzienlijke verschillen zien tussen de volume-susceptibiliteiten van deeltjes met dezelfde buitendiameter. Ook tonen we aan dat het gebruik van AC magnetische velden leidt tot een lager wordende sus-ceptibiliteit voor toenemende veldfrequentie. Tenslotte presenteren we een chipontwerp waarmee de hydrodynamische verstoring ten gevolge van het oppervlak verkleind kan worden, door grotere stroomdraden te gebruiken,
waartussen het deeltje heen en weer kan bewegen op enige micrometers afstand van het chipoppervlak.
In Brownse-bewegingsanalyse vangen we deeltjes van 150 tot 450 nm in een instelbare magnetische potentiaalput boven een geïntegreerde stroom-draad. Het histogram van de verdeling van de deeltjesposities geeft dan de grootte van de deeltjesmagnetisatie weer. Met behulp van rechte stroomdra-den zien we verschillen in volumesusceptibiliteit van meer dan een factor twee voor hetzelfde deeltjestype. Door het gebruik van rechte stroomdraden in combinatie met barrières op het oppervlak, of stroomdraden met een inge-snoerde vorm, verkleinen we de meetfout in de susceptibiliteitsbepaling tot minder dan 10%. We laten ook zien dat ingesnoerde stroomdraden gecombi-neerd kunnen worden met een extern uniform veld, waardoor meer informatie verkregen kan worden over deeltjeseigenschappen zoals anisotropie en per-manent moment.
Het tweede deel van dit proefschrift beschrijft een studie naar de dynamica van deeltjes die biologisch gebonden zijn aan een oppervlak. We laten zien dat een optisch evanescent veld gebruikt kan worden om de thermische variatie in de hoogte van gebonden deeltjes te meten met nanometerresolutie, omdat de intensiteit van het licht verstrooid door de deeltjes afhangt van de deeltjes-hoogte boven het oppervlak. Doordat we de deeltjes binden met een molecuul van bekende lengte (dsDNA van 290 bp/99 nm) kunnen we aantonen dat hoogtevariaties inderdaad kwantitatief bepaald kunnen worden. We laten zien dat de nauwkeurigheid van de hoogtebepaling bepaald wordt door deeltjesei-genschappen zoals vorm, gladheid, en inhoud. Optimale resultaten zijn gevon-den voor polystyrene en magnetische deeltjes die glad en rond zijn.
Vervolgens laten we zien dat we de driedimensionale thermische beweeg-lijkheid van het deeltje kunnen gebruiken om de binding tussen het deeltje en het oppervlak te karakteriseren. Als modelbinding gebruiken we vier verschil-lende DNA lengtes (590 bp/201 nm, 290 bp/99 nm, 141 bp/48 nm en 105 bp/36 nm). Grafieken van de minimale hoogte, gemiddelde hoogte en maxima-le hoogte als functie van de positie in het vlak weerspiegemaxima-len de verschilmaxima-len in bindingslengte, bindingsflexibiliteit en bindingsoriëntatie. Voor verschillende ensembles van deeltjes, gebonden aan het oppervlak met een bepaalde DNA lengte, laten we zien dat de beweeglijkheid in hoogte maximaal gelijk is aan de bindingslengte, maar we zien grote verschillen tussen deeltjes, die we toe-schrijven aan niet-specifieke interacties tussen het deeltje en het oppervlak.
De beweeglijkheid van een gebonden deeltje kan worden beïnvloed door aangelegde krachten. We laten zien dat een magnetische gradientkracht ge-richt naar het oppervlak toe de gebonden deeltjes dichterbij het oppervlak brengt, maar dat een gradient kracht gericht van het oppervlak af de gebonden deeltjes niet altijd verder van het oppervlak af brengt. Dit kan worden ver-klaard door magnetische anisotropie in de deeltjes, wat leidt tot uitlijning in het veld. We beschrijven een model dat laat zien hoe de uitlijning gebonden
deeltjes in gedefinieerde driedimensionale posities brengt, met mogelijke gereduceerde hoogtes. Hoewel uitlijning tegenstrijdig kan zijn met de gewens-te magnetische manipulatie, kan het ook extra informatie geven over bijvoor-beeld rotationele vrijheid van de binding en bindingsflexibiliteit.
Tot slot laten we zien dat een gebonden deeltje ook hoger boven het op-pervlak gebracht kan worden door een elektrostatische kracht, geïnduceerd door de vervanging van de buffer door een buffer van lagere ionische sterkte. We laten zien dat de verandering in deeltjeshoogte afhangt van zowel de bindingslengte als de ionische sterkte van de buffer, en geven een kwantitatief model om de gemeten hoogteveranderingen voor deeltjes gebonden met verschillende DNA lengtes te verklaren.
De verbeterde kennis over de magnetische eigenschappen van individuele deeltjes en over de beweeglijkheid van individuele deeltjes gebonden aan een biosensoroppervlak, volgend uit de experimenten beschreven in dit proef-schrift, kan leiden tot lagere detectielimieten en hogere specificiteit in toekom-stige biosensoren.
reface
When we look into our Universe, containing billions of galaxies, stars and planets, we are astonished by its enormous dimensions. But when we look into our own human body, a similarly large and complex Universe can be found. Millions of cells (~1014) work together intelligently and even the system of one
cell is a complete world in itself with thousands of components, many of which are still not fully understood.
For our outer Universe, science and technology keep developing newer and more sophisticated instruments to visualize far away stars and galaxies that cannot be seen by the naked eye. Similarly, for our inner Universe, more and more technology is developed to dive deeply into the mysteries of our existence. For astronomy, visualization starts with a telescope, for biology the first step is the optical microscope.
With an optical microscope we can examine, for example, body tissues or body fluids, such as blood or saliva. The resolution of a traditional optical microscope is limited by the diffraction of light; for visible light on the order of a few hundred nanometers. Human cell diameters range from a few microme-ters (blood platelet, 2 to 3 μm, granule cell of the cerebellum, 4 by 4.5 μm) to around 135 μm (anterior horn cell of the spine) and can thus well be studied in an optical microscope. Even structures inside the cell, like the nucleus, the cytoskeleton or the mitochondria, can be distinguished.
The discovery of fluorescent labels further expanded the possibilities of optical microscopy. In the 1960s and 1970s, a protein was isolated from a jellyfish, which exhibits bright green fluorescence when exposed to blue light. In the 1990’s the first proof was shown of the cloning of the green fluorescent
protein and the application of the protein as a tool for biologists. Fluorescent molecules can be specifically coupled to certain cell structures or molecules and then be excited by light with the appropriate wavelength. By filtering out the wavelength emitted by the fluorescent molecule one can visualize the presence of these molecules, and thereby of the desired structures.
The use of fluorescent labels also opened the possibility to study objects smaller than the diffraction limit for visible light, determined by the so-called Airy disk size of a few hundred nanometers. Because only the light emitted by the fluorescent molecule is filtered out, even the presence of one fluorescent label within an Airy disk, and thus of a molecule of interest, can be detected. Moreover, its position can be determined with a few nanometers precision by a center of mass analysis of the intensity profile. However, two molecules within the detection area can still not be distinguished. This can for example be solved by diluting the sample such that the probability of having more than two molecules in the imaging region is small.
In the 1990’s also another technique was introduced to detect and obtain information on a single-molecule level, namely to couple individual molecules to larger objects of tens of nanometers up to a few micrometers, for example metal particles, polystyrene particles, quantum dots or magnetic particles. These objects can in turn be more easily visualized in optical microscopy. Moreover, they can also be manipulated using for example pipettes and suc-tion, optical tweezers (focused laser beams) or magnetic forces, giving the possibility to obtain additional information on the properties of the molecules of interest.
During the years of “space travelling” in the inner Universe of the human body, we have learned that the mechanisms of the body are extremely com-plex, but that the body is very well designed. Continuously, cells are being disintegrated and regenerated, damage to the body in the form of cuts, frac-tures or bruises is healed, and the immune system constantly fights external attacks from bacteria and viruses. However, there are also many possibilities for failure, leading to complications and diseases. Hundreds of years ago, diseases could only be diagnosed by visible symptoms and verbal descriptions. These days, we have many techniques to look inside the body non-destructively, for example by measuring signals from the body in electroe-ncephalography (EEG) or electrocardiography (ECG), or by imaging the inside of the body with for example radiography or magnetic resonance imaging (MRI). When the techniques improved to study the human body down to a molecular level, it became clear that many diseases could also be diagnosed by detecting the presence of certain molecules in the body.
This thesis will describe research on nanoparticles and microparticles, used as labels for biological molecules, and visualized in optical microscopy, for the so-called fields of bionanotechnology and biomolecular diagnostics.
Chapter 1 Introduction
Chapter
1
Introduction
1.1 Biomolecular Diagnostics
Many diseases can be diagnosed by detecting the presence and/or proper-ties of certain molecules in body fluids.1,2 For example, the presence of the
protein ‘prostrate-specific antigen’ in blood may point to prostate cancer and an elevated concentration of the protein ‘troponin’ in blood can indicate dam-age to the heart muscle. To detect the concentration of a certain target mole-cule in a fluid, this molemole-cule has to be distinguished from other substances in the fluid (i.e. other molecules, cells) by some specific property. For example, the target molecules may specifically interact with an enzyme, or the target molecules can be separated from other substances by binding them specifically to a surface. In case the target molecule is a protein, capture on a surface can be done using the specific reaction between an antibody and its antigen, the target molecule. Antibodies are proteins, made by the body’s immune system, that are composed of basic structural units and have a small region at the tip of the protein (‘epitope’) that is extremely variable and can bind to a specific molecule. The target molecules can also be specifically coupled to a label, for example a fluorescent molecule, to facilitate detection of the target molecule in a complex biological mixture. The total sequence of steps required to
deter-mine the concentration of target molecules in a sample is called an assay. The test is called an immunoassay when antibodies are used to capture the target molecules.
Biomolecular assays are mostly performed in centralized laboratories by specially trained people. Some tests, however, can be carried out in doctor’s offices or even in home settings with devices that are called point-of-care biosensors. Examples of point-of-care biosensors are the glucose sensor that diabetics use to measure the concentration of glucose in their blood, and the home pregnancy test that detects markers for pregnancy in urine. When more point-of-care biosensors become available for specific diseases, medical diag-nostics and treatment can be accelerated, simplified and improved. In case of for example damage to the heart muscle, early detection can even save lives.
1.2 Point-of-Care Biosensors
A point-of-care biosensor is a small, preferably hand-held, device that can detect the presence of a certain molecule in a body fluid, see Fig. 1.1. Detection can take place in many different ways, such as optical, piezoelectric or electro-chemical detection. The blood glucose sensor, for example, detects glucose by measuring the electric current that is induced when the enzyme glucose oxi-dase converts glucose.3 In the home pregnancy test urine flows along a strip
where it mixes first with nanoparticles coated with antibodies that can specifi-cally bind to pregnancy markers and then flows over a detection area contain-ing also antibodies that bind specifically to pregnancy markers.1,4 A change of
color of the detection zone, which can be detected visually, points to the pres-ence of nanoparticles on the detection zone and thus to the prespres-ence of preg-nancy markers in the urine.
Fig. 1.1. Example of a point-of-care biosensor that detects the presence and/or concentra-tion of a certain molecule, for example glucose, in blood.
To make a successful point-of-care biosensor, the device has to detect the analyte fast, in seconds to minutes, and in a small amount of body fluid, such as a drop of blood from a finger prick. It has to be sensitive and specific, as many other molecules will also be present in the body fluid, and it should give a reliable diagnosis with hardly any false negatives or false positives. It should also be easy to operate and be robust against external circumstances like temperature.
For molecules that are present in body fluids in high concentrations, like glucose having concentrations of millimolar in blood (~1 gram/liter), exten-sive research has been done, and products have been brought to the market that fulfill the requirements for point-of-care diagnostics. It is however very challenging to meet all the requirements for molecules with lower concentra-tions. For example, the concentration of troponin in blood is on the order of femtomolar to picomolar, a factor 109-1012 lower than the glucose
concentra-tion.5 Present clinical tests in hospital laboratories can detect a few picomolar
of troponin,6 and current tests in a preclinical phase can detect concentrations
down to 100 femtomolar.5 However, large equipment and specially trained
people are still needed.
To detect low concentrations of target molecules in a biosensor, the use of particles (tens of nanometers up to a few micrometers) to label the molecules becomes advantageous, as the size of these particles enables easier manipula-tion and/or detecmanipula-tion.1,7 In practical applications, such particles are often
called ‘beads’, as their shape is usually spherical. The particle labels can be coated with antibodies so they can specifically capture the molecules of inter-est from the body fluid. A common format to apply particle labels for the detection of target molecules is a surface sandwich format, see Fig. 1.2. In a sandwich assay, the target molecule is sandwiched between an antibody on the surface and an antibody on the label. The label can subsequently be detected, and the concentration of labels on the surface is a measure for the concentra-tion of target molecules in the sample.
The detection sensitivity and operation speed of a biosensor are influenced by the number of target molecules that can be brought to the detection area in a certain amount of time. Without labels the speed of this process is limited by diffusion, which is usually slow. For example, a protein with a diffusion coeffi-cient on the order of 10-10 m2/s takes on average more than 2 hours to diffuse
over 1 mm. Some improvement in speed can be gained by using for example fluid flows. However, to get to very low concentrations of target molecules, it is desirable to fish the target molecules out of the fluid and quickly transport them to the detection area. A way to do this effectively is by using magnetic particle labels. Magnetic particles can easily be manipulated using magnetic fields and magnetic field gradients, enabling efficient movement of labels through the fluid to capture the target molecules and subsequent transport of labels and target molecules to the detection area.
target molecule label antibody 1 antibody 2 other molecule other molecule
Fig. 1.2. Sandwich assay format, in which an antibody on the surface binds to a specific target molecule. On the other side of the target molecule a label is attached via a second antibody.
1.3 Magnetic Particles
Magnetic particles have been commercially available for many years and are widely used in laboratories to extract desired biological components, such as cells, organelles or DNA, from a fluid.8,9 Magnetic particles are adsorbed to the
biological component, and with a magnetic field the particles, with attached biological components, are brought close together, after which the fluid and remaining content can be removed or replaced.
Magnetic particles can have any size from a few nanometers to a few micrometers and can contain magnetic materials such as iron, nickel, cobalt, neodymium-iron-boron, samarium-cobalt or magnetite.9 Often the particles
are coated with polymers to prevent the formation of aggregates and to facili-tate biological functionalization. Nanometer sized particles (5–50 nm) are usually composed of a single magnetic core with a polymer shell around it. Larger particles (30 nm – 10 μm) can be composed of multiple magnetic cores inside a polymer matrix. A special class of magnetic particles are the so-called superparamagnetic beads. Superparamagnetic beads are composed of thou-sands of very small magnetic grains, generally between 5 and 15 nm, embed-ded in a polymer matrix, see Fig. 1.3. When magnetic grains are so small, their magnetic moment can randomly flip between different orientations due to the thermal energy, and therefore the grains have no net magnetic moment. How-ever, when a magnetic field is applied, a competition is induced between the magnetic energy and the thermal energy, resulting in a net magnetization. The magnetization is per volume unit larger than that of paramagnetic materials, hence the name superparamagnetism. The magnetization is, however, lower than that of permanent magnetic particles, but problems with aggregation are avoided due to the absence of a magnetic moment without a field.
50 nm
Fig. 1.3. Transmission Electron Microscopy image of a superparamagnetic bead (300 nm Ademtech bead) consisting of 5 to 15 nm iron oxide grains (black) in a polystyrene matrix (grey).
In the last decade extensive research has been done on the use of magnetic particles for a novel generation of biosensors. Magnetic particles can be used for efficient transport, for faster assay kinetics, for improving binding specific-ity and as labels for detection. Examples of studies on the first three aspects will be described below, and detection will be treated in the next section (Section 1.4).
The transport of magnetic particles in microfluidic systems or biosensors is investigated in several ways, such as using mechanically moving permanent magnets, sets of electromagnets with specific actuation schemes, or micropat-terned and integrated current wires.10,11 Some examples in the latter category
are the use of two dimensional wire matrices12 and planar coils.13,14
Wirix-Speetjens et al used for example a set of two alternately actuated tapered wires to transport particles along the wires.15 A combination of external magnets and
integrated structures was used by Gunnarson et al, who demonstrated the transport of magnetic particles with micronsized permalloy elements and rotating magnetic fields.16
The improvement of assay kinetics using magnetic particles has been described by Baudry et al, who showed that the association rate of binding can be increased by bringing particles close together in chains using magnetic fields.17 Arrays of dense pillars of magnetic particles can be used to efficiently
trap biological material flowing through a microfluidic channel.10 Dittmer et al
show how the efficiency of target molecule capture and binding to the surface is increased approximately five-fold in a biosensor when magnetic actuation of magnetic particle labels is used instead of diffusion and sedimentation.18 To
optimally use magnetic particles to improve assay kinetics, it should be noted that the size of the particles is important. Small particles have a high diffusivity (rotational, translational), which increases the rate of finding reactive sites.17,19
However, large particles are generally more magnetic and thus easier to ma-nipulate or detect. For different types of biosensors, particle diameters be-tween 50 nm and 3 μm have been used.20
A technique to improve binding specificity using magnetic particles was developed by Lee et al, who showed that in a surface-immunoassay involving
magnetic particles, a magnetic force can be used to remove magnetic particles that are settled onto the surface by gravity or by weak non-specific interac-tions.21 Particles that are specifically bound to the surface via a target molecule
and antibodies remain at the surface, because the specific binding force is higher than the applied magnetic force.
1.4 Detection of Magnetic Particles
Detection of magnetic particles in a biosensor can take place in many different ways. We categorize the different assay types used for detection into surface assays, cluster assays and particle assays and the different detection tech-niques into magnetic detection and optical detection, as illustrated in Fig. 1.4. This section will shortly address several examples of each category. The re-search presented in this thesis is done in the framework of biosensors using a spin-valve GMR sensor (surface assay, magnetic detection) and an evanescent field (surface assay, optical detection). These two examples will be further explained. Surface Assay Cluster Assay Particle Assay Magnetic Detection Optical Detection
Fig. 1.4. Schematic illustration of the detection of magnetic particles in biosensors using three types of assays and two different detection techniques.
In a surface assay magnetic particles are bound via a target molecule to a sensor surface, for example in a sandwich format as explained in Fig. 1.2. The magnetization of the bound particles can be detected with a magnetic sensor and is a measure for the number of target molecules. Luxton et al for example bind magnetic particles to a polyester disk (~15 mm in diameter), remove unbound magnetic particle from the disk area using magnets and then detect the bound particles using a flat spiral pick-up coil (~4 mm in diameter), which is mainly sensitive to the disk area.22 Nikitin et al used porous filters (a few
millimeters in size) with a large surface area to bind magnetic particles that are flowing through the filter and detect the particles by a frequency depend-ent coil measuremdepend-ent based on the non-linear field-dependence of the particle magnetization.23
on-detect magnetic particles near a sensor surface. Baselt et al first demonstrated the use of a Giant Magnetoresistance (GMR) sensor in a biosensor in 1998.24
The GMR sensor gives a resistance change when exposed to a magnetic field from magnetic particles. Recently, Osterfeld et al (2008) showed for example the successful application of a GMR sensor to simultaneously detect several potential cancer markers at sub-picomolar concentration levels.25 Also other
magnetic sensors that give a resistance change when exposed to a magnetic field, have been investigated for biosensing, such as Magnetic Tunnel Junction (MTJ) sensors and Anisotropic Magnetoresistance (AMR) sensors.20 Slightly
different operation mechanisms are found in a Hall-sensor, in which a mag-netic field leads to an asymmetric charge density distribution and thus an electric potential,26-28 and a Giant Magneto Impedance (GMI) sensor, which is
based on the change in impedance of a conductor passed by a high-frequency current when exposed to a magnetic field.29
Besides detecting magnetic particles by measuring the particle magnetiza-tion, also the time-dependent relaxation behavior of the magnetization can be used in a technique called relaxometry.30 The time-dependent particle
mag-netization is influenced by both the required time for physical (Brownian) rotation of the particle (facilitated by thermal energy), and the required time for rotation of the magnetic moment (Néel) inside the particle. Kotitz et al showed for example that particles that are bound to a surface can no longer physically rotate.31 When the Néel relaxation is long (on the order of seconds),
the bound particles exhibit a magnetic moment even after removal of a mag-netic field, which can be detected using a Superconducting Quantum Interfer-ence Device (SQUID).
Optical detection of magnetic particles in a surface assay is possible by imaging of the sensor surface. Mulvaney et al showed that single particles can be counted on a typical 100 μm microarray spot using bright field imaging with a microscope objective and a CCD camera.32 Morozov et al showed optical
detection of magnetic particles using dark field imaging,33 and Bruls et al
showed optical detection of magnetic particles that are illuminated using an evanescent field.34
The second assay type is a cluster assay, in which particles are bound to other particles via target molecules (see Fig. 1.4). Kotitz et al showed in a relaxometry experiment how clustering of the particles leads to considerable inhibition of the physical (Brownian) rotation.35 When the magnetic field is
removed, the Néel relaxation of the bound particles, on the order of 100 milli-seconds, leads to a characteristic relaxation behavior over a period of about a second, which can be measured using a SQUID. The shape of the measured curve reveals the amount of clusters and thus the concentration of target molecules. Optically, clusters can be detected using various techniques. Baudry
et al showed for example a change in intensity between the transmitted and
are formed due to binding of a target.17 Physical rotation of particle clusters
induced by rotating magnetic fields can further improve optical detection as is shown by Petkus et al36 and proposed by Ranzoni et al.37
The last assay type is a particle assay, in which the capture of a target molecule by a magnetic particle is detected directly. This is possible by tech-niques that rely on the hydrodynamic radius of a particle, which is dependent on the particle size and is thus influenced by target binding. Connolly et al proposed to measure the relaxation frequency of particles in a changing mag-netic field (Hz-kHz).38 When the dominating magnetic relaxation mechanism is
physical (Brownian) rotation, which depends on the hydrodynamic radius, target binding changes the dynamic magnetization behavior of the parti-cles.39,40 Optically, magnetic relaxation can also be detected using the Faraday
effect, in which the light polarization depends on the magnetization of the particle.41 For particles that are non-spherical,42 or partly coated with metal, 43-45 physical rotation by a rotating magnetic field also leads to fluctuating optical
signals.
It should be noted that next to magnetic and optical detection of magnetic particles in the different assay types, also other detection techniques are possible. For example, Lee et al report the use of a cluster assay combined with Nuclear Magnetic Resonance (NMR) detection, where magnetic particle clus-ters are detected by the change in spin-relaxation time of the surrounding water molecules, measured by planar microcoils.46 Mak et al showed a surface
assay combined with electric detection, in which magnetic particles, bound to microelectrodes, inhibit an electrochemical reaction and thus lead to a reduc-tion in the measured electrochemical current.47 Also, detection techniques
exist that use magnetic particles primarily as carriers for performing particle or surface assays and combine this with additional labels for detection, such as subsequent electrochemical detection of an enzyme reaction,48 or optical
detection of fluorescent molecule labels in an evanescent field.49
a) b)
sensor
electrical signal current magnetic field lines
sensor
N S N S
Fig. 1.5. a) Detection of a magnetic particle, bound to the surface in a sandwich format, using an integrated current wire to magnetize the particle and a spin-valve sensor to measure the in-plane component of the particle magnetization. b) Microscope image of a biosensor chip containing four sensor units. Each sensor unit consists of two current wires
Now, we will further elaborate on the two cases that form the basis for the work in this thesis, namely detection of magnetic particles in a surface assay using either a spin-valve GMR sensor or an optical evanescent field. A magnetic biosensor based on a spin-valve GMR sensor is shown in Fig. 1.5.18,50,51 A
spin-valve sensor gives an electrical resistance change when exposed to a magnetic field. Magnetic particles are bound to the sensor surface, for example in a sandwich format, and are magnetized by integrated current wires in the chip surface. Because the sensor is mostly sensitive to in-plane magnetic fields, the out-of-plane magnetic field caused by the current wire gives no signal change. The magnetized particle generates a magnetic dipole field, which does have an in-plane component, and therefore leads to a signal change in the sensor. To increase the signal to noise ratio of the sensor signal, the actuation and sensing currents are modulated at high frequencies (~MHz). Using 500 nm magnetic particles, Dittmer et al demonstrated the detection of 10 pM of parathyroid hormone in 5 minutes,18 and Koets et al demonstrated the detection of 4 pM
DNA amplicons in 3 minutes.52
The detection of magnetic particles using an optical evanescent field is shown in Fig. 1.6.34 A collimated beam of light from a light-emitting diode
irradiates the sensor surface in a condition of total internal reflection, generat-ing an evanescent optical field that penetrates into the fluid by only a sub-wavelength distance. An object that is present in the evanescent field with a refractive index larger than that of the fluid, for example a magnetic particle, will scatter part of the light and thereby frustrate the total internal reflection. This leads to a reduction in the intensity of the reflected light, which can be measured using a CCD camera. The geometric conditions of the detection
upper magnet fluid LED 2D camera lower magnet tape
Fig. 1.6. Optical detection of magnetic particles in an evanescent field, using the principle of frustrated total internal reflection. Particles present in the evanescent field scatter part of the light, leading to a reduction in intensity signal of the reflected beam, which is measured by a camera. Magnets are used to manipulate the particles in the assay
system permit electromagnets to be situated directly above and below the fluid chamber, thus enabling magnetic manipulation of the magnetic particles. In Fig. 1.6 the lower magnet attracts the particles toward the sensor surface and the upper magnet pulls unbound particles away from the surface. Using 500 nm magnetic particles, Bruls et al demonstrated the detection of sub-picomolar troponin concentrations in a few minutes.34
1.5 Research Goal
The research described in this thesis is done in the framework of improv-ing the sensitivity of biosensors that use magnetic particles in surface assays. For a decreasing concentration of target molecules, the concentration of mag-netic particles at the sensor surface also decreases. The detection limit of a biosensor is determined by a number of factors, such as the sensitivity of the sensor to the detection label, the inherent physical noise in the detection signal, the noise in the detection signal induced by components in the biologi-cal fluid and the specificity of the binding process. In this thesis we focus on a better understanding of two aspects of magnetic particles in a biosensor that may lead to improved detection limits in future biosensors, namely a precise determination of the magnetic properties of the particles, and a specific char-acterization of the bond between the magnetic particle and the sensor surface. In both cases we take a single-particle approach, because the properties and behavior of particles of one type may vary. Below, we will give a few examples of aspects that may be improved by a better understanding of particle proper-ties and binding characteristics.
Well-defined magnetic properties are important for controlled transport in biosensors, for efficient and uniform removal of unbound particles and for accurate detection using magnetic detection techniques. The accuracy in measuring the number of particles on a sensor surface using magnetic detec-tion is determined by the accuracy in the signal per label. Magnetic sensors have been shown to have a good sensitivity as several tens of particles can easily be detected; with detection limits even down to a few particles50 or
single particles.20,53,54 However, when the magnetic properties of the used
magnetic particles vary, the sensor signal cannot always be correctly related to the number of magnetic particles. For example, when the variation in magneti-zation between particles is 50%, over 100 particles are required to reduce the error in detection signal to 5% (the counting error reduces with the square root of the number of particles). Because the particle magnetization is ex-pected to be mostly dependent on the volume of the particle, variations of 50% in particle magnetization may already occur due to 15% variations in particle radius. Commercially available batches of particles exist wherein particle sizes vary more than 100%.
Efficient removal of particles that are not specifically bound to the surface also requires uniform magnetic properties of the particles. Particles may be
non-specifically bound to the surface in different ways, such as shown in Fig. 1.7. For certain assays, it has been shown that the binding force of a specifically bound particle is higher than that of a non-specifically bound particle (see Section 1.3 and Lee et al21). Applying a magnetic force can then remove
non-specifically bound particles from the surface while non-specifically bound particles remain at the surface. This increases the specificity of the assay and thus the sensitivity of a biosensor. It is also possible that non-specifically bound parti-cles are attached stronger to the surface than specifically bound partiparti-cles,21 or
that both weaker and stronger non-specific binding occurs simultaneously, in which case more complex procedures are needed to distinguish between specific and non-specific binding. For all force differentiation techniques it is essential that the applied force is well-defined and uniform for all particles on the sensor surface, so that the distinction is made on the bond force between particle and surface and not on particle properties.
target molecule particle antibody 2 antibody 1 non-specific specific ?
Fig. 1.7. Several examples of particles that are non-specifically bound to a sensor surface, compared to a particle that is specifically bound to the surface in a sandwich immunoas-say.
The second topic, characterization of the bond between a magnetic particle and the sensor surface, is important for a better understanding of the nature of non-specific binding, which can lead to improved techniques to avoid or re-move non-specifically bound particles. Precise bond characterization can also lead to the development of techniques to make a distinction between specifi-cally bound particles and non-specifispecifi-cally bound particles in an integrated biosensor without the need to remove non-specifically bound particles. Inte-grated detection techniques for bond characterization may also open the road to the development of so-called functional biosensors,55 where not only the
presence of a target molecule is detected, but also information can be obtained on the physical and/or chemical properties of the target molecule. In this thesis we investigate whether we can use the thermal mobility of a bound particle as a way to characterize the bond.
Both subjects, the magnetic properties of particles and the mobility of bound particles, are investigated by studying the dynamics of individual parti-cles near a biosensor surface. The movement of individual partiparti-cles is analyzed using an optical microscope and a camera. We focus on the use of
superpara-magnetic beads (see Section 1.3) between 300 nm and 1 μm, as these sizes have been shown to be suitable for biosensors using magnetic actuation and detection with a spin-valve sensor17 or an evanescent field.18,34
The magnetic properties of individual particles are investigated using two techniques: on chip magnetophoretic analysis and on-chip confined Brownian motion analysis. Section 1.6 gives an overview of magnetic characterization techniques and existing studies on magnetophoresis. Section 1.7 describes studies that use the confined Brownian motion of particles to determine forces and energies. Existing techniques to characterize molecules or bonds by the movement of bound particles are described in Section 1.8. Finally, Section 1.9 describes techniques to study the out-of-plane movement of particles using optical microscopy.
1.6 Magnetic Particle Characterization
The goal of the first part of this thesis is to measure the magnetic properties of particles between 300 nm and 1 μm on a single-particle level. The properties of magnetic particles depend on a variety of factors, such as the type and amount of magnetic material, the shape of the particle and the microstructure, which can give rise to strong differences between individual magnetic particles (as will be further explained in Section 2). In this thesis we focus mainly on the measurement of the particle susceptibility, which is the ratio between the particle magnetization and the applied magnetic field. Typical magnetic fields used in biosensors with integrated current wires are fields between 1 and 10 mT.20,56,57 Therefore, we aim to measure the particle susceptibility in fields
below 10 mT. In this section we give an overview on existing techniques to characterize magnetic particles, suitable either for ensembles of magnetic particles or for individual particles, and we pay particular attention to the technique of magnetophoretic analysis, which we use in this thesis. An over-view of the different characterization techniques is given in Table 1.1.
Large ensembles of magnetic particles can be characterized using the well-known magnetometers ‘Superconducting Quantum Interference Device’ (SQUID), where the current in a superconducting loop is influenced by the presence of a magnetic field, or a ‘Vibrating Sample Magnetometer’ (VSM), where a sample is physically vibrated, giving a varying magnetic flux that induces a voltage change in pick-up coils.58-61 In a SQUID or a VSM the sample
magnetization can be measured without applying a magnetic field or as func-tion of an applied field. Both the low-field susceptibility and the saturafunc-tion magnetization of magnetic particles can thus be determined. Moreover, from the shape of the magnetization curve, information can be obtained on the grain size distribution, which influences the particle magnetization.59 Also indirect
techniques can be used to study ensembles of magnetic particles, for example infrared spectroscopy, which can detect vacancies in crystal structures that can
Table 1.1. Overview of different techniques to measure the magnetic properties of ensem-bles of magnetic particles and the magnetic properties of individual magnetic particles.
Ensemble Individual particles
Magnetization/ susceptibility VSM SQUID Magnetophoresis Hall sensor SQUID
Confined Brownian motion analysis*
High frequency susceptibility
Pick-up coils
Toroidal coil impedance Short-circuit transmission line
Particle rotation (high frequency) Internal grain structure VSM / SQUID Infrared spectroscopy Mössbauer spectroscopy
Particle rotation (high frequency)
Hall sensor
Remanent moment VSM / SQUID Magnetophoresis
Particle rotation (low frequency)
* see Section 1.7.
lead to varying magnetic properties,9 or Mössbauer spectroscopy, which gives
information about the directions of the atomic moments in nanoparticles by measuring the average hyperfine fields that influence the internal magnetic order.9,59
In addition to studying the static magnetization of an ensemble of magnetic particles, also the dynamic magnetization of the ensemble can be determined. The dynamic susceptibility of a particle ensemble is measured using a mag-netic excitation field with a certain frequency (Hz to GHz) and detection tech-niques that measure the time-varying magnetic flux of the sample, such as pick-up coils, toroidal coil impedance measurements or a short-circuit trans-mission line technique.62-66
The most important technique to study the magnetic properties of
individ-ual particles is magnetophoretic analysis. In magnetophoretic analysis
infor-mation on the magnetic properties of particles is obtained by measuring the speed of the particles induced by a magnetic field gradient (‘magnetophoretic mobility’). This technique was described in 1960 by Gill et al, who used a microscope and a stopwatch to measure the speed of red blood cells (10μm large) and 5 to 50 μm polystyrene particles in a magnetic field gradient.67
Automated video analysis measurements came up in the 1980s and a well-defined set-up was designed by Reddy et al in 1996, who measured the suscep-tibility of beads of 4.5 to 13.5 μm in a known field gradient.68 The years
there-after, effort was made to create defined and largely constant gradients with magnetic set-ups suitable for use in a microscope environment, for example by
Chalmers et al.69 Nowadays, magnetophoretic analysis is the most common
technique for analyzing individual particles or magnetically labeled biomate-rial such as cells.70,71 The set-ups are mainly suitable for measuring in high
magnetic fields (200 mT to 2.3 T) and for using magnetic particles larger than 4 μm.68,72-74
A variety of slightly smaller magnetic particles of around 1 to 4 μm was studied in a low field of 22.5 mT by Häfeli et al.75 They compared the variations
in properties between 16 types of particles, and found that the magnetopho-retic mobility of different types of particles was not related to their saturation magnetization. This shows the importance of studying the magnetic properties at lower fields. Häfeli et al attributed the differences between high field mag-netization and low field magmag-netization to variations in the internal distribution and porosity between the different particle types. Differences within one type of particles could not be determined accurately, especially not for the 1 μm size range where the error in magnetophoretic mobility determination was roughly 50%.
The results of Häfeli et al showed several problems that can occur when analyzing small particles, for example particle speeds that are lower than the displacement resolution of the imaging system and the thermal or Brownian motion of the particles. Brownian motion is the irregular, random motion that small particles exhibit in a fluid due to their thermal energy, which originates from collisions of the particles with molecules in the fluid. Häfeli et al showed that Brownian motion can cause particles to move out of the focus plane of the microscope, which prevents following them for a longer time. Brownian mo-tion also influences the accuracy of the speed determinamo-tion in magnetopho-retic analysis, as was already described by Gill et al in 1960.67 They stated that
in order to overcome this error for small particles, either high forces are required (by using high field gradients) or the time period of the measurement should be long.
Another technique to study the magnetization of individual particles was proposed by Mihajlovic et al, who used a phase-sensitive micro-Hall magne-tometer, with a size on the order of the particle diameter, to map the suscepti-bility of an individual 1.2 μm superparamagnetic bead in fields between 1 mT and 100 mT.76 Additionally, they derived the size distribution of the grains
inside the bead from fitting the magnetization curve to a theoretical distribu-tion. SQUID, mentioned before as a technique to characterize particle ensem-bles, can also be used to study the magnetization of individual particles, as was shown by Yamaguchi et al.77
In addition to studying the susceptibility or induced (low-field) magnetiza-tion of magnetic particles, also research has been done on studying the perma-nent moment of magnetic particles. Shevkoplyas et al show using both SQUID measurements on ensembles of particles and magnetophoretic analysis on individual particles that a permanent moment can give a significant
contribu-tion to the particle magnetizacontribu-tion at low fields.56 It has also been shown that
magnetic particles can be physically rotated in a rotating magnetic field.45,78
Romano et al proposed to measure the magnetic torque on a particle in a rotating field by determining the phase lag between the particle and the field.79
They trapped a single superparamagnetic bead in an optical tweezer and used optical asymmetries (impurities on the bead surface) to detect particle rota-tion. Particle rotation was investigated in more detail by Janssen et al, who found that the origin of the magnetic torque and the rotational behavior of a superparamagnetic bead was a small permanent moment of up to 1% of the saturation magnetization.55 Additionally, Janssen et al studied the rotation
behavior at fields with high (up to 3 MHz) frequencies, from which they could determine the characteristic relaxation times of the grains inside the bead, which relates to the grain size distribution.
Finally, characterization of magnetic particles can also be done by first separating a large ensemble into smaller batches with more similar magnetic properties. Magnetic particle separation is investigated in both macroscopic and miniaturized devices.10 Usually, the separation is based on a fluid flow
through a region with a high magnetic field gradient that either retains or deflects magnetic particles based on their magnetic properties.10,80 The
mag-netic field gradients can be created with permanent magnets, external elec-tromagnets, integrated current wires or micromagnets on a chip.10,80-82 Most
magnetic separation techniques are suitable for particles above the microme-ter diamemicrome-ter range due to the high field gradients needed for smaller particles. However, it has recently been shown that separation is also possible for parti-cles in the nanometer range (50-700 nm), but the variation in size between the separated batches is still quite significant.83,84
1.7 Confined Brownian Motion Analysis
In the previous section we showed that one of the main errors in measuring magnetic properties with magnetophoretic analysis is the Brownian motion of the particles. In this thesis we describe a method that uses the Brownian motion of small magnetic particles, with a diameter between 100 nm and 1 μm, to characterize their individual magnetic properties.
When an external force acts on a particle that exhibits Brownian motion, the motion is random, but biased. The trajectory of particle that experiences a restoring force to a certain position, i.e. is captured in a potential energy well, may look like Fig. 1.8a. Three techniques have been developed to obtain infor-mation on the external force from the measured trajectory.85 The first method
examines the histogram of the measured particle positions, see Fig. 1.8b. The probability density to find a bead at a certain position is given by a Boltzmann factor that relates the potential energy of the particle to the thermal energy. The second method assumes that the potential well is harmonic and the restor-ing force linear and uses the equipartition theorem to link the mean squared
displacement of the position data to the spring constant of the potential well. The third method also assumes a harmonic potential well and uses the Fourier transform of the position data to determine both the diffusion coefficient of the particle as well as the spring constant. This is called the power spectrum method. time po s it ion a) position co un ts b)
Fig. 1.8. a) Trajectory of a particle undergoing Brownian motion in a potential well with a restoring force to the centre. b) Histogram of the particle positions in the well.
In this thesis we use confined Brownian motion analysis combined with the probability density method to determine the magnetic energy of a particle captured in a magnetic potential well. The magnetic energy is directly related to the magnetization of the particle, i.e. the more strongly magnetic the particle is, the deeper the magnetic energy potential well experienced by the particle and the less freedom of motion the particle has due to thermal energy. Below we will describe a number of experimental approaches in which confined Brownian motion analysis is used, namely in optical tweezers, in particle-surface interaction measurements, and in measurements of magnetic forces.
Confined Brownian motion analysis came up in the 1980s with the discov-ery of optical tweezers by Ashkin et al.86 In an optical tweezer small dielectric
particles are trapped by a focused laser beam, which gives a restoring force to the centre of the beam that is usually modeled by a linear relation correspond-ing to a harmonic potential well. Confined Brownian motion analysis is used to determine the forces and energy of the tweezer. A recent study by Wei et al shows for example the mapping of the 3-dimensional optical force field on silica microparticles of 2.6 μm in an optical tweezer, using both the probability density method and the power spectrum method.87
Confined Brownian motion analysis is also used in studying the interaction between a particle and a surface in a fluid. The interaction consists of a van der Waals interaction (due to interacting atomic dipoles) and an electrostatic interaction (due to surface charges on the particle and the surface). Prieve et al (1986) proposed to determine the potential energy profile between a particle and a surface by measuring the distribution in separation distances and apply-ing the probability density method.88 First, they determined the separation
distance indirectly via the reduced translational speed of a particle close to a surface88 and later they developed a technique called Total Internal Reflection
Microscopy to measure the separation distance directly89 (see also Section 1.9
for more information on Total Internal Reflection Microscopy).
Confined Brownian motion analysis has also been applied to magnetic potential wells. For example, Gosse et al measured the magnetic force on 4.5 μm magnetic particles trapped in a magnetic tweezer, consisting of elec-tromagnetic coils, using both the equipartition theorem method and the power spectrum method.90 Mirowski et al used the equipartition theorem to
charac-terize the forces of magnetic tweezers, consisting of micrometer sized per-malloy elements, on 2-3 μm magnetic particles.91 Helseth et al investigated
localized magnetic field gradients near the interface of a magnet, for example around a domain wall, using 2.8 μm magnetic particles and the probability density method.92 These studies used optical detection of the particles and
video analysis to determine particle positions.
One study was published on characterizing the magnetic properties of particles using confined Brownian motion analysis, simultaneously to the work described in this thesis. Blickle et al characterized the susceptibility of 2.7 μm and 4.5 μm superparamagnetic beads by trapping a single superparamagnetic bead near a surface using an optical tweezer.93 The interaction energy between
the particle and the surface was determined using the probability density method. Subsequently, the particle was slightly pulled out of the optical trap using an external electromagnet and the magnetic force on the particle was calculated by the measured change in interaction energy. From the magnetic force, the particle magnetization was determined in fields ranging from about 8 mT to 200 mT.
1.8 Bond Characterization
The goal of the second part of this thesis is to investigate whether the mobility of a bound particle can be measured as a way to obtain information on the bond (see Section 1.5). In this thesis we combine the technique of tethered particle motion with the technique of magnetic tweezers to analyze the bond-complex between a particle and the surface. This section will give a short introduction on the use of particle labels to study single molecules in both optical tweezer and tethered particle motion experiments, and presents exam-ples of the use of magnetic particles and magnetic tweezers for both single-molecule research and bond-characterization studies. Next to the detection of the in-plane mobility of bound particles, we detect also the out-of-plane mobil-ity, which will be further described in the next section (Section 1.9).
The use of particle labels to characterize single molecules was initiated around the 1990s in the so-called field of ‘single-molecule biophysics’ by the emergence of two important milestones. The first milestone was the discovery of optical tweezers by Ashkin in 1986 (see also Section 1.8).86 When a single
