1

MgO MTJ biosensors for immunomagnetic lateral-flow

detection

Ricardo Jorge Penelas Janeiro

Dissertação para obtenção do Grau de Mestre em

Engenharia Física Tecnológica

Júri

Presidente: Prof. Dr. João Carlos Carvalho de Sá Seixas

Orientador: Prof. Dr. Susana Isabel Pinheiro Cardoso de Freitas

Vogal: Prof. Dr. Paulo Jorge Peixeiro de Freitas

Outubro de 2010

ii

iii

Acknowledgement

Com chegada ao fim da tese de mestrado, etapa final do curso, impõe-se o dever e felicidade

de agradecer a todos os que permitiram empreender toda esta jornada académica. Dever porque sem

todos eles quem sabe o que poderia ter acontecido pelo caminho ardiloso que é a vida de um jovem

universitário; Dever porque de todos eles o meu ser se formou, por causa da interacção com cada um

se moldou. Alegria porque poder agradecer significa não só o concluir de um ciclo, como a existência

de a quem agradecer, sinal de um caminho não percorrido isoladamente.

Começo por agradecer à instituição que permitiu a realização deste trabalho, o INESC-MN,

que me proporcionou o meu primeiro contacto com a Ciência: a verdadeira ciência crivada de

dificuldades, imprevistos e cujos resultados nem sempre premeiam o esforço dispendido, mas que é

recompensadora quando é atingido um objectivo. Agradeço ao Professor Paulo Freitas e à

Professora Susana Freitas pela oportunidade de trabalhar no INESC-MN e pela orientação científica

e experimental proporcionada. Um agradecimento também ao Dr. Ricardo Ferreira pela paciência e

disponibilidade com encarou todas as infindáveis questões. Obrigado também ao resto da equipa

científica do INES-MN, dos seus engenheiros aos alunos pela disponibilidade e ajudas prestadas

sempre que solicitados. Um especial obrigado aos alunos de mestrado de curso de física pelos

excepcionais momentos de companheirismo (Cláudio, Raquel, Zita).

Prosseguindo na enumeração daqueles que marcaram a etapa que agora termina é

imperioso referir o conjunto de amigos com quem partilhei diariamente a experiência académica no

IST (made me a man). Desde o primeiro grupo com que travei conhecimento / já conhecia – Mike,

Ana, Romão – às mais recentes crianças que incautamente entraram neste curso – Patos e Amigos

Lmta – passando pelo mais experiente e já Mestre José Gustavo Rebelo, e pelo caríssimo D. Diogo

Capelo, até aquela que pertence a uma classe só dela - Raquel (obrigado coisa :) ) o meu muito

obrigado por tudo sendo que muitas vezes o proporcionarem uma gargalhada foi o suficiente e ideal.

Agradeço a toda a minha família por todo o apoio e confiança depositada em mim e nas

minhas capacidades (pobres ingénuos :p). Um muito obrigado aos meus pais, elementos

fundamentais e sempre presentes em qualquer momento.

Não me escudando na minha confessa e evidente inaptidão para a “escrita livre e pessoal”

desde já as minhas desculpas às inúmeras pessoas que aqui ficaram por referir.

Obrigado Mãe, Obrigado Pai, Obrigado Raquel =)

Ricardo Janeiro

Dezembro de 2010

iv

v

Resumo

Nos tempos que correm, é evidente o potencial das nanotecnologias, as quais são capazes

de fornecer os meios necessários para a detecção e rastreamento de patogéneos na forma de

sensores de tamanho micrométrico muito sensíveis.

O objectivo deste trabalho é a fabricação de sensores magnetoresistivos, com resposta linear

capazes de executarem testes de reconhecimento através da detecção de nanopartículas

funcionalizadas, magneticamente polarizadas e previamente ligadas a biomoléculas. Os sensores

magnetoresistivos usados foram Junções de Efeito de Tunel, as quais apresentam por norma um

valor elevado de magnetoresistência. Neste trabalho junções individuais foram ligadas em série,

formando vectores de 360 elementos individuais. Tal foi feito com o propósito de ganhar na

detectividade dos dispositivos.

Várias amostras foram processadas e magnetoresistências da ordem de 70% foram

conseguidas, assim como sensibilidades de 1.24%/Oe na região linear.

Palavras chave: Sensores magnetoresistivos de efeito de túnel, junções de efeito de túnel

de MgO, ruído, fluxo lateral, séries de junções de efeito de túnel.

vi

vii

Abstract

In the modern times is clear the potential of the nanotechnologies, which are capable of

providing the necessary means for the detection and screening of pathogens in the form of very

sensitive micron size sensors.

The aim of this work is the fabrication of magnetoresistive (MR) sensors with linear response,

capable of performing tests of recognition by the detection of functionalized nanoparticles magnetically

polarized which had been previously linked to biomolecules. The magnetoresistive sensors used were

Magnetic Tunnel Junctions which present a high value of magnetoresistance. In addition, in this work

individual Magnetic Tunnel Junctions were connected in series, forming arrays of three hundred and

sixty individual elements. Such was done with the purpose of gain in the devices’ detectivity.

Several samples were patterned and magnetoresistances of the order of 70% were achieved,

as well as sensitivities of 1.24%/Oe in the linear range.

Keywords: Tunneling magnetoresistance (TMR) sensors, Mgo MTJ, noise, lateral

flow, MTJ series.

viii

ix

Contents

Acknowledgement ................................................................................................................................... iii

Resumo ....................................................................................................................................................v

Abstract................................................................................................................................................... vii

Figure List ................................................................................................................................................ xi

Tables List ............................................................................................................................................. xiii

1. Introduction ...................................................................................................................................... 1

1.1. Thesis’s framework – motivation ............................................................................................. 1

2. Theoretical background – Magnetism ............................................................................................. 3

2.1. Magnetic field and Magnetization ............................................................................................ 3

2.2. Magnetic Materials Classification ............................................................................................ 4

2.3. Micromagnetism for ferromagnetic material ............................................................................ 7

2.3.1. Exchange energy ................................................................................................................. 7

2.3.2. Magnetocrystalline Anisotropy – Anisotropic Energy .......................................................... 7

2.3.3. Shape anisotropy – Demagnetizing Energy ........................................................................ 8

2.3.4. Zeeman Energy ................................................................................................................. 10

2.3.5. Interlayer coupling forces .................................................................................................. 10

2.4. Magnetoresistance ................................................................................................................ 11

2.4.1. Anisotropic Magnetoresistance (AMR) .............................................................................. 11

2.4.3. Tunneling Magnetoresistance Effect ................................................................................. 15

3. MgO MTJ biosensors for immunomagnetic lateral-flow detection – System and sensor design .. 23

3.1. Reader system design ........................................................................................................... 23

3.2. Sensor design ........................................................................................................................ 25

3.2.1. Noise and detectivity implications .................................................................................. 26

3.2.2. Application: theoretical preview ..................................................................................... 28

4. Micro fabrication – Systems, techniques and processes protocols............................................... 29

4.1. Sputtering Deposition Systems: Magnetron .......................................................................... 29

4.1.1. Nordiko 2000 ................................................................................................................. 31

4.1.2. Nordiko 7000 ................................................................................................................. 32

4.1.3. Ultra High Vacuum-UHVII – Oxide Sputtering System.................................................. 33

4.2. Ion Beam Systems ................................................................................................................ 34

4.3. Pattern Transfer - Techniques and Processes ...................................................................... 36

x

4.3.1. Pattern Transfer Techniques ......................................................................................... 36

Etching ........................................................................................................................................... 36

Lift-off ............................................................................................................................................. 36

4.3.2. Direct Write Laser Optical Lithography (DWL) .............................................................. 37

4.4. Annealing setup ..................................................................................................................... 40

5. Characterization methods – Measurement techniques ................................................................. 43

5.1. Vibrating Sample Magnetometer (VSM) ................................................................................ 43

5.2. Manual Transport Measurement Setup ................................................................................. 43

5.3. Profilometer ........................................................................................................................... 45

5.4. Ellipsometer ........................................................................................................................... 46

6. Microfabrication – Process steps................................................................................................... 47

6.1. Stack deposition .................................................................................................................... 47

6.2. Bottom contact definition ....................................................................................................... 48

6.3. MTJ pillar definition ................................................................................................................ 50

6.4. Pillar lateral isolation .............................................................................................................. 51

6.5. Top contact definition – Metallization .................................................................................... 51

6.6. Sample insulation .................................................................................................................. 52

6.7. Sample dicing ........................................................................................................................ 52

6.8. Annealing ............................................................................................................................... 52

6.9. Encapsulation ........................................................................................................................ 52

7. Results ........................................................................................................................................... 53

8. Conclusions and Future work ........................................................................................................ 63

Bibliography ........................................................................................................................................... 65

A. Appendix ........................................................................................................................................ 67

B. Appendix ........................................................................................................................................ 69

Run Sheet for MTJs Array fabrication: Influenza_ 2010 ................................................................... 69

xi

Figure List

Figure 1.1: Lateral flow recognition. ........................................................................................................ 2

Figure 2.1: Diamagnetic material ............................................................................................................ 4

Figure 2.2: Paramagnetic material .......................................................................................................... 4

Figure 2.3: Ferromagnetic material ......................................................................................................... 5

Figure 2.4: Antiferromagnetic material .................................................................................................... 6

Figure 2.5: Ferrimagnetic material .......................................................................................................... 6

Figure 2.6: Magnetization and demagnetizing field ................................................................................. 8

Figure 2.7: a) single domain; b) 2 domains; c) 4 domains. ..................................................................... 9

Figure 2.8: Domain Wall .......................................................................................................................... 9

Figure 2.9: Néel coupling....................................................................................................................... 10

Figure 2.10: Oscillatory character of RKKY coupling ............................................................................ 11

Figure 2.11: Transfer curve of a GMR multilayer. ................................................................................. 12

Figure 2.12: Spin valve structure ........................................................................................................... 13

Figure 2.13: Scheme of electron scattering in a GMR device. .............................................................. 14

Figure 2.14: Magnetic tunnel jucntion structure. ................................................................................... 15

Figure 2.15: Couplings in the MTJ stack. .............................................................................................. 15

Figure 2.16: Tunnelling phenomenon .................................................................................................... 16

Figure 2.17: Scheme of a MTJ composed by a pinned layer and a free layer with parallel anisotropies

............................................................................................................................................................... 18

Figure 2.18: MTJ transport behaviour to the parallel anisotropy. .......................................................... 19

Figure 2.19: Scheme of a MTJ composed by a pinned layer and a free layer with crossed anisotropies

............................................................................................................................................................... 20

Figure 2.20: MTJ transport behaviour to the crossed anisotropy: linear response. .............................. 21

Figure 3.1: Scheme- Magnetoresistive lateral flow detection – wheel concept. ................................... 23

Figure 3.2: Wheel device prototype ....................................................................................................... 24

Figure 3.3: Mask used in this work ........................................................................................................ 25

Figure 3.4: MTJ series: array of MTJs; each individual element has a 30x2µm2 area. ........................ 26

Figure 3.5: 3D scheme of the MTJs array’s. ......................................................................................... 26

Figure 3.6: Expected output: particles magnetized vertically. ............................................................... 28

Figure 4.1: Scheme of a magnetron system. ........................................................................................ 29

Figure 4.2: Relation between the wear of a target and the magnetic field lines. .................................. 30

Figure 4.3: Nordiko 2000 deposition camber schematic. ...................................................................... 31

Figure 4.4: Nordiko 7000 - disposition of the all process modules, dealer and load-lock. .................... 32

Figure 4.5: Schematic side view of UHVII. ............................................................................................ 33

Figure 4.6: Picture of UHVII................................................................................................................... 34

Figure 4.7: Schematic view of a IBD system: Z configuration. .............................................................. 35

Figure 4.8: Etching process. .................................................................................................................. 36

Figure 4.9: Lift-off process ..................................................................................................................... 36

Figure 4.10: Picture of the SVG tracks. ................................................................................................. 38

xii

Figure 4.11: Optical lithography – photo-resist mask creation. ............................................................. 39

Figure 4.12: Figure of the DWL system. ................................................................................................ 39

Figure 4.13: Picture of the used annealing setup. ................................................................................. 40

Figure 4.14: Temperature cycle in the annealing process. ................................................................... 41

Figure 5.1: VSM system. ....................................................................................................................... 43

Figure 5.2: Electrical scheme for measurements with 2 (on the left) and four (the right one) probes. . 44

Figure 5.3: INESC-MN profilometer system .......................................................................................... 45

Figure 5.4: ............................................................................................. 45

Figure 5.5: Picture of the ellipsometer. .................................................................................................. 46

Figure 6.1: Standard MTJ stack used ................................................................................................... 47

Figure 6.2: Resulting photoresist after the 1st lithography revelation. ................................................... 48

Figure 6.3: Bottom contact definition: resulting structure after the ion milling. ...................................... 48

Figure 6.4: Picture of bottom contact. ................................................................................................... 49

Figure 6.5: Left - Wet bench’s picture: ultrasound machine and hot bath; Right - microstrip solution. . 49

Figure 6.6: Pillar junction definition. Left - After the 2nd

lithography; Right – After the 2nd

etch ............ 50

Figure 6.7: Picture of the pillar junction. ................................................................................................ 50

Figure 6.8: MTJ pillar and bottom contact insulation ............................................................................. 51

Figure 6.9: Top contact definition ......................................................................................................... 51

Figure 6.10: Picture of the top electrode. .............................................................................................. 52

Figure 6.11: Sample patterned and mounted on a chip-carrier............................................................. 52

Figure 7.1: Close view of one MTJ series with the 4 contacts identified as a, b, c and d. .................... 53

Figure 7.2: ZarMTJ1 transfer curve ....................................................................................................... 53

Figure 7.3: ZarMTJ2 transfer curve. ...................................................................................................... 54

Figure 7.4: ZarMTJ3 transfer curve ....................................................................................................... 54

Figure 7.5: ZarMTJ5 transfer curve. ...................................................................................................... 54

Figure 7.6: ZarMTJ1 and ZarMTJ3; measurement for 360 elements. .................................................. 55

Figure 7.7: The RxA of the sensors in ZarMTJ1 and ZarMTJ2 samples. ............................................. 55

Figure 7.8: ZarMTJ1 - TMR vs number of junctions. ............................................................................. 56

Figure 7.9: Transfer curve between contacts of two independent series. ............................................. 57

Figure 7.10: N2#1 transfer curve. .......................................................................................................... 58

Figure 7.11: N2#2 transfer curve. .......................................................................................................... 58

Figure 7.12: N2#4 transfer curve. .......................................................................................................... 58

Figure 7.13: N2#1 - TMR vs number of junctions ................................................................................. 59

Figure 7.14: TMR vs Vbias ...................................................................................................................... 60

Figure 7.15: TMR of a MTJ decreasing with the applied voltage .......................................................... 61

Figure 7.16: Graph - Response of one sensor under magneto nanoparticles excitation. ..................... 62

Figure .A.1: Mask used .......................................................................................................................... 67

Figure A.2: Top: Series detail ................................................................................................................ 68

Figure A.3: 3D scheme of MTJ’s connected in serious ......................................................................... 68

xiii

Tables List

Table 4.1: Deposition setpoints of Nordiko 2000. ................................................................................. 31

Table 4.2: Standard conditions operation in the used modules ............................................................ 33

Table 4.3: Deposition conditions of UHV2. ............................................................................................ 34

Table 4.4: Ion milling used set values for assist gun of both IBD systems. .......................................... 35

Table 7.1: Variations in the stack of samples with Si substrates. ......................................................... 53

Table 7.2: Variations in the stack of samples with glass substrates. .................................................... 58

xiv

1

1. Introduction

The last century saw the advent of non-classical science, beyond the limits of the Newtian

physics and Maxwell’s electromagnetism. These scientific progresses were instrumental in the

development of several technological areas, having allowed, among others, major advances in

electronics, whose importance has assumed a dominant role nearly in all the areas of human

intervention. In this context of technological progression a pre-requisite for the success of any system:

it’s contained area and volume. From this requirement born microelectronics. driving the need for a

continued miniaturization; Currently, it is no longer just the need of space that dictate the importance

of micro / nanotechnologies, which has extended from their traditional applications’ area in computer

systems and automated control systems, to the most recent biosciences. A family of small-scale

devices that is particularly interesting is the magnetoresistive family, whose applications include

sensors for read heads in hard disk drives, memories, biosensors and diagnostic techniques using

brain mapping.

In fact, in biosciences context, which are currently showing themselves as a scientific area

with major scientific and social impact, it’s clear the potential of the magnetoresistive (MR) sensors, as

a way of providing the necessary means for the detection and screening of health threatening

pathogens. Was in order an application under the biorecognition and detection that this work was

developed, using the advantages of a particular type of magnetoresistive sensors: the magnetic tunnel

junctions (MTJ’s) sensors – high spatial resolution provided by their small size, and higher sensitivities

and output values comparing with the others MR sensors.

1.1. Thesis’s framework – motivation

Done in the INESC-MN facilities, with special emphasis in the INESC clean-room, the work

done in this thesis is integrated in an international project which has, as ultimate goal, to develop a

new diagnostic tool for influenza virus detection based on an immunochromatographic assay with

functionalized magnetic nanoparticles as markers of the virus and ultra sensitive field transducers for

their detection and quantification.

Actually, the present methods for human influenza virus based on cell cultures have a

response time frame of some days, too long to help physicians to make clinically relevant decisions

during the first days of the disease, when the treatment will be more effective. There are already some

immnunochromatography based tests, which yield a response in less than 30 minutes, but they suffer

from low sensitivity (only 50 to 70% of true influenza cases are detected).

2

These rapid immunochromatographic tests are fast assays based on the immunological

capture of a ligand attached to an active microparticle when it flows over a solid support where an anti-

ligand has been immobilized.

The chosen approach for this project takes advantage of the exploitation of magnetic

properties of magnetic nanoparticles as active particle, which will be detected and quantified by ultra-

sensitive magnetic sensors. The first developments in field sensors lead to inductive readers, in which

an electrical response is induced in some coils placed around the field source: if a magnetic flux

variation through the coils exists, an electromotive force is created, like stated in Faraday’s law. The

coils method has a low sensitivity so in the project the coils are replaced by magnetoresistance based

sensors (magnetoresistive sensors).

Resuming, the final goal of the project is the creation of an integrated platform including a

system that enables the processing of samples allegedly virus contaminated and the system of

detection and quantification.

The working principle is like follows: first the testing sample containing the virus is put in

contact with magnetic nanoparticles functionalized with a specific antibody to the particular virus

strain. The fluid is then put in contact with a porous membrane strip; the fluid sample passes through

the membrane until the virus attached to the magnetic markers are recognized by membrane

immobilized antibodies; finally the presence of a particular virus strain is detected through the sensing

of the magnetic fringe field created by the bound magnetic nanoparticles, using a magnetic sensor. Is

worthy to note that this concept can be used to detect other virus than influenza one.

Figure 1.1: Lateral flow recognition. a) the fluid sample is put into contact with magnetic particles; b)virus is attached to nanoparticles antivirus; c) fluid sample migrates through the membrane strip.

3

2. Theoretical background – Magnetism [1],[2],[3],[12],[14],[17]

The magnetic field and magnetic properties of materials play a major role in the behaviour of

devices that were fabricated and studied during this thesis. Therefore this first section will introduce

and remember some basic concepts and definitions.

2.1. Magnetic field and Magnetization

A magnetic field is a field of force that arises from two sources: electric currents (generically

electric charge moving) and materials with an intrinsic nonzero magnetic moment (magnetic

materials). In fact, the magnetism of the matter is due to the movement of electrons from atoms, which

are responsible by the microscopic magnetization currents postulated by Ampère before knowing the

structure of matter, so we can see the common ground in these two sources of field: it results from

electric charge movement.

Because of the nonexistence of magnetic charges every magnetic field source has a dipolar

form (the first term appearing in the dipolar treatment of a distribution of magnetic sources is the

dipolar), and so, magnetic materials, being a magnetic field source, have a magnetic moment

associated just like current loops. Through the dipolar moment is possible to define the magnetization

of a material:

m is the magnetic moment of the material and V its volume.

The two types of sources contributing to the magnetic field can be related by the following

expression:

, µ0 free space permeability

with H taking into consideration the free currents influence, while B considers all the currents ( free

and magnetization currents; so B takes into account the magnetization of the medium).

The influence of the H field over the magnetization of magnetic materials is the information

contained in magnetic susceptibility:

4

2.2. Magnetic Materials Classification

Diamagnetism

The atoms of diamegnetic materials do not present a permanent magnetic moment in absence

of applied magnetic field, since they don’t have unpaired electrons on their completely filled orbital

shells. However, when a magnetic field is applied a small intensity response occurs: these materials

have a negative small susceptibility, so they tend to align the magnetic moment of their atoms against

the applied field.

The diamagnetic behaviour is present in every material, but due to its small magnitude is often

negligible or covered by stronger magnetic effects.

Paramagnetism

In the absence of a magnetic field, the atoms of a paramagnetic material have a permanent

magnetic moment due to the unpaired electrons on their partially filled shell. Since these atoms, as

magnetic dipoles, poorly interact with one another they get a random orientation due to thermal

agitation. In the presence of an external magnetic field the moments will gradually align along with it

as the intensity of the field increases. So paramagnetic materials have a positive, but small

susceptibility.

So, for both diamagnetic and paragmetic materials, when a magnetic field is applied a

magnetization proportional to that field is produced:

and so the magnetic field is:

Figure 2.1: Diamagnetic material: Left - In absent of applied field; Right - Under applied field.

Figure 2.2: Paramagnetic material in field absence (left) and under its influence (right).

5

Diamagnetic materials:

- ;

- is thermally independent;

- has a very small absolute value;

Paramagnetic materials:

- ;

- is generally small (however greater than the absolute values of diamagnetic materials): for

low temperatures when the thermal agitation decreases significantly paramagnetic materials

exhibit higher values.

Ferromagnetism

Ferromagnetic materials exhibit a large permanent magnetization even when a magnetic field

is not present. Like in paramagnetism, the atoms of ferromagnetic materials have unpaired electrons.

The main difference between these two types of magnetism lies on the stronger interaction existing

between ferromagnetic atoms – the exchange interaction. If a magnetic field rises, the individual

magnetic moments tend to align with the field, so the susceptibility is positive and typically large.

Above a certain temperature, the Curie temperature, ferromagnetic materials behave like

paramagnetic ones.

Figure 2.3: Ferromagnetic material – moments align without an applied field.

6

Antiferromagnetism

In antiferromagnetic materials the exchange interaction between their atoms results in

individual magnetic moments ordered in such way that moments of consecutive atoms in the lattice

are aligned anti-parallel. Since in these two sub-lattices the moments have an equal magnetic

magnitude, the net magnetic moment in the absence o field is zero. This kind of materials has a small

and positive susceptibility.

Antiferromagnetic materials behave like paramagnetic ones above the Néel

temperature (it is the analogous of the Curie temperature in ferromagnetism); each

ferromagnetic also has a characteristic blocking temperature above which the strong

interlayer exchange interaction vanishes and an antiferromagnetic layer loses its “ability” to

pin a ferromagnetic layer.

Ferrimagnetism

The magnetic moments exhibit alternate orientation like in ferromagnetism. However because

there is more than one type of magnetic ion in the material, not all magnetic moments have the same

absolute value. This means that they have a non null net magnetization.

Figure 2.4: Antiferromagnetic material – antiparallel alignment of the moments; no net magnetization.

Figure 2.5: Ferrimagnetic material

7

2.3. Micromagnetism for ferromagnetic material

Ferromagnetic and antiferromagnetic materials are key elements in the behaviour of the

devices focused by thesis. So to better understand their properties it is a good approach to make

some considerations about the interactions and the associated energies that occur in or with these

materials.

2.3.1. Exchange energy

The spontaneous ordering of ferromagnetic atomic moments is mainly due to the exchange

interaction of their atoms, which is independent of the total magnetic moment of the sample.

According to the theoretical atomic dipoles model for ferromagnets, the main difference

between ferromagnets and paramagnets is that the each permanent dipole of the ferromagnets

interact strongly with their nearest neighbours dipoles. The interaction energy between two sequenced

atoms is described by the Heisenberg model:

is the coupling constant between the spins of the neighbour atoms, and is usually considered

constant through a material and Si and Sj are the spins of each neighbour.

The total exchange energy of a material implies sum over all pairs of nearest neighbours;

generalizing to the continue approximation the exchange energy of entire crystalline lattice is

a is the lattice constant.

If the magnetization varies too rapidly in a short range the exchange energy will be very high,

so in order to keep it in a minimum value the magnetic dipoles have a preference to remain aligned

with each other.

2.3.2. Magnetocrystalline Anisotropy – Anisotropic Energy

Crystalline materials are magnetically anisotropic because there is a preferential direction for

orientation of the dipoles along the main crystallographic axis of the structure. This means that exist a

sort of internal field which forces the net magnetization to align with certain axis of the crystalline

structure, defining the so called easy axis or easy direction.

The magnetocrystalline anisotropy energy, Ek, is the work needed to rotate the sample

magnetization to a certain direction out of the easy axis direction, and is calculated as follows:

8

where K is an energy density, and α is the angle between the magnetization and the easy axis.

Is clear to see that this energy will have a minimum for α=0, which means that the

minimization of this energy implies the magnetization to prefer align with the easy axis. This

contributes to the “memory effect” of ferromagnets (after the magnetic field get off the material will

tend to align their dipole with the easy axis).

2.3.3. Shape anisotropy – Demagnetizing Energy

The demagnetizing energy corresponds to the interaction between the magnetization of the

material and the demagnetizing field, Hd. His expression follows below:

In order to understand the appearance of the demagnetizing field a good strategy could be to

establish the analogy with the electric field. A polarized material creates a distribution of electric

charges at the surfaces whose normal has the same direction of the polarization vector P. That

charges are responsible for the rise of an electric field/electrostatic potential (Poisson’s equation). With

the magnetization a similar phenomenon happens: on the surfaces, at the material boundaries (those

that are perpendicular to some magnetization component) the magnetization is no longer continuous,

and “magnetic charges” appear. These charges are the sources of the demagnetizing field that will

oppose to the normal magnetization of the material; this means that this energy tries to become the

magnetization parallel to “charged” surfaces.

Minimizing the demagnetizing energy corresponds to rotate the magnetic atomic dipoles of the

sample, so that the number of charges created on surfaces is minimal. That makes the material

become divided into different magnetic domains oriented in reverse directions. This way, the magnetic

charges formed by one domain will cancel the charges of neighbour domains, reducing Hd and Ed.

Figure 2.6: Magnetization and demagnetizing field: the field is greater the closer are the “magnetic charges”.

9

If the demagnetizing energy was the only one ruling the behaviour of ferromagnetic materials

they would divide themselves in as many domains as necessary to completely eliminate the

demagnetizing field resulting in a zero total magnetization.

However, indeed, the exchange energy avoids this to happen, because it has the inverse

effect: its minimization cause the atomic moments of neighbours to align with each others. So, it is the

balance of these two energies is responsible by the creation of magnetic domains separated by

domain walls, which are regions where the orientation of the magnetic dipoles that belong to them is

not constant, rotating from the orientation in the first domain to the orientation of the second (Figure

2.8: Domain WallFigure 2.8). The thickness of these domain walls is also influenced by the anisotropy

energy: this energy forces the moments to align by the easy axis, so it would “prefer” no walls at all,

but the exchange energy would prefer a wall so big as the material so that neighbour dipoles would be

as parallel as possible; so anisotropy energy will try to create a thin wall and exchange energy will try

to enlarge it.

It is important to note that wasn’t written that there are magnetic monopoles. Isolated magnetic

charges doesn’t exist ( ), so that charges referred above always appears in pairs constituting a

dipole. In fact if one tries to split the north pole from the south pole of a magnet will end with two

smaller magnets, each with one north and one south pole.

Figure 2.7: a) single domain; b) 2 domains; c) 4 domains.

Figure 2.8: Domain Wall

10

2.3.4. Zeeman Energy

There is one last energy term that describes the magnetic properties of a ferromagnet: the

Zeeman energy. The Zeeman energy is related to the interaction between the magnetization of a

material and the external applied field (Ha) and is given by:

This energy represents the amount of work necessary to rotate the magnetization by a certain

angle with respect to the applied field, and would be a minimal value when the magnetization became

align along the magnetic field direction.

2.3.5. Interlayer coupling forces

2.3.5.1. Néel Coupling

It is right to say that the Néel coupling is present in every ferromagnetic interlayer

discontinuity, because it is associated with roughness of the ferromagnetic interfaces (it is reasonable

to say that exist always some level of roughness in an interface). The magnetostatic interactions

between the “magnetic charges” at the ferromagnetic interface cause a ferromagnetic coupling.

The Neel energy is given by:

2.3.5.2. RKKY Coupling

Ruderman-Kittel-Kasuya-Yodsida coupling or indirect oscillatory exchange interaction is an

exchange interaction between two ferromagnetic layers separated by a non magnetic spacer. The

wave functions of the atoms of the two ferromagnetic layers interact with each other in a way mediated

by the spacer, whose thickness is responsible for the oscillatory character of the interaction between a

ferromagnetic and an anti-ferromagnetic coupling.

Figure 2.9: Néel coupling.

11

2.4. Magnetoresistance

Magnetoresistance is the property of a material to change the value of its electrical resistance

when the value of the applied magnetic field changes. The magnitude of this effect can be expressed

numerically as follows and is presented as a percentage:

2.4.1. Anisotropic Magnetoresistance (AMR)

The AMR effect consists in the change of electrical resistivity with the orientation of

magnetization in respect with the direction of the electrical current in the material. Thus, the resistivity

of a magnetic material depends on the angle between the magnetization and the current direction and

can be described by the following equation:

For most materials the resistance is maximized when the current is parallel to the

magnetization and minimized when it is perpendicular.

To measure the amplitude of this effect (to know how much the resistivity varies) is necessary

measure the maximum and the minimum, so it is necessary to saturate the magnetization in parallel

and perpendicular directions. With these two values it’s possible to calculate the anisotropic

magnetoresistivity ratio:

Typically AMR is somewhere between 2 a 6%.

Figure 2.10: Oscillatory character of RKKY coupling with the spacer thickness. The strength of the ferromagnetic and anti-ferromagnetic coupling decreases with the spacer thickness.

12

2.4.2. Giant Magnetoresistance (GMR)

Giant magnetoresistance is an effect observed in thin film structures composed of alternating

ferromagnetic and non magnetic layers. The observed effect is a significant change in the electrical

resistance of such structures depending on whether the magnetization of adjacent ferromagnetic

layers are parallel or anti-parallel aligned. The overall resistance is low for parallel alignment and high

for anti-parallel alignment. Therefore the intensity of the GMR effect is given by:

Multilayer GMR

In this structure two or more ferromagnetic layers are separated by a non-ferromagnetic

spacer (FM/AFM/FM structure - e.g. Fe/Cr/Fe – the first GMR sample). To a certain thickness of the

spacer, the RKKY coupling between the adjacent ferromagnetic layers becomes anti-ferromagnetic,

making it, in the absence of a magnetic field, energetically preferable for the magnetizations of

adjacent ferromagnetic layers to align anti-parallel. When an external filed is applied and reaches a

certain value which makes energetically favourable to break the inter layer coupling the ferromagnets

will align parallel to each other and parallel with the field. This rotation and alignment of the

ferromagnetic layers are followed by a drop in the electrical resistance of the structure. In the

multilayer configuration with 10 or more stacks MRs of about 10% were achieved.

Figure 2.11: Transfer curve of a GMR multilayer. The red arrows represent the magnetization of each layer to different applied fields.

13

Spin Valve

In spin valve GMR two ferromagnetic layers are separated by a thin non-magnetic metallic

layer spacer, and an antiferromagnetic material is contiguous to one of the ferromagnetic layers (the

pinned layer).

In the spin valve configuration the electrical current flows in plane with the layers: Current In

Plane - CIP geometry.

The antiferromagnetic material is an exchange bias, which sets an exchange anisotropy at the

interface with the ferromagnet by the exchange interlayer coupling. This way the anti-ferromagnetic

material sets and pines the magnetization of the ferromagnetic material providing a fixed reference.

The other ferromagnetic layer is free to rotate with the applied field and align along it. With this

configuration the spin valve resistance has a maximum value when the magnetization of the free layer

and the pinned layer are anti-parallel and a minimum value when they are parallel.

It is the spin-dependent conductance in the ferromagnetic layers that supports the GMR effect.

This spin dependence can be explained by the so called “two channel model”. This model states that

the spin-flip scattering process that takes place in ferromagnets have a much larger time-scale than all

other processes contributing to electrical resistance. Thus it is assumed that the electrons conserve

their spin, and that there are two distinct families of electrons, which are distinguished by their state of

spin and can be considered independently. In GMR devices most of the current will flow through the

non-magnetic material, which is usually the layer with the lowest resistivity. In a non-magnetic material

the population of conduction electrons, the ones with energy close to Fermi level, is constituted by an

equal number of electrons with spin up and with spin down. Since the scattering of electrons is mostly

dominated by collisions between electrons of the same spin is possible to understand that in nom-

magnetic materials there is no spin orientation granted with a higher scattering rate. This is so while

the electrons are confined to the non-magnetic material, but when they approach the interface

ferromagnet/non-ferromagnet that is no longer true because the density of states at the Fermi level is

asymmetric with respected to the spin. This means that according to the ferromagnet magnetization,

near the Fermi level the number of spins aligned parallel with the magnetization direction will be higher

Figure 2.12: Spin valve structure; the current flows in the layer’s plane.

14

than the number of spins with the anti-parallel alignment. As result, and considering, as already

mentioned, that the scattering is dominated by collisions between electrons with the same spin, when

the electrons enter the ferromagnet the scattering probability will increase for one of the spin directions

and decreased for the other, depending on the magnetization direction. When the magnetization of the

two layers are anti-parallel the two spins orientations are scatter at the same rate (scattering of one of

the spin orientations is increased in one of the ferromagnet and the other spin orientation scattering

increase in the other ferromagnet; the same applies to the complementary spin orientation in each

ferromagnets, which the scattering rate decreases). If the magnetizations of both ferromagnetic layers

are parallel, the GMR device structure will no longer be symmetric with respect to spin orientation.

Now the two ferromagnetic layers will have a larger population of one of the spin types (the same type

for both ferromagnets) in the vicinity of the Fermi level depending again on the orientation of the

magnetization. The scattering for this particular spin type will be strongly decreased and its

contribution for the current flowing through the device will be very high. To the opposed spin direction

the scattering will be increased and its contribution for the current very small.

The typical MR signal of a spin valve reaches the 9%.

To a better understand this model the relation between the band structure of a three layer

ferromagnetic / anti-ferromagnetic / ferromagnetic device and electron scattering it can be seen the

figure bellow, which have at the right side the equivalent resistor model (Figure 2.13).

Figure 2.13: Scheme of electron scattering in a GMR device.

15

2.4.3. Tunneling Magnetoresistance Effect

Magnetic Tunnel Junction (MTJ)

A magnetic tunnel junction is a structure constituted by two ferromagnetic layers separated by

an insulator layer thin enough to allow the electrons tunneling. In the MTJ’s structure the current flows

perpendicularly to the layer’s plane (current perpendicular to plane geometry – CPP geometry).

Figure 2.14: Magnetic tunnel junction structure; the current flows perpendicularly to the planes.

Like in the spin valve configuration one of the ferromagnetic layers is pinned, providing a fixed

reference while the other is free to move under an external field application.

In the MTJs used in this work the pinned layer role is performed by a synthetic anti-

ferromagnetic (SAF) structure, which is composed by two ferromagnetic layers separated by non-

magnetic one. With this SAF structure is created an alternative way to have a pinned layer to the

simple anti-ferromagnet / ferromagnet coupling. Hence, the anti-ferromagnetic layer is a exchange

bias material and pines the Ferromagnet Layer 1 which pins the Ferromagnet Layer 2 by RKKY

coupling, which competes with the ferromagnetic Neel coupling, several orders of magnitude inferior to

RKKY coupling ( Figure 2.15). This way a stronger pining is achieved than using just the inter layer

exchange coupling.

Figure 2.15: Couplings in the MTJ stack.

16

Being strictly forbidden in classical physics, the tunnelling phenomenon rises from the double

corpuscular-wave character of the electrons. The wave-like behaviour raises a non-null probability of

the electrons cross a trough an energetically forbidden barrier (the oxide insulator in the MTJ

structure).

When no voltage is applied (V=0) between the two ferromagnets (the free and the pinned

layer), the Fermi level of the two electrodes is the same and no current flows. If a voltage is applied

between the electrodes the energy levels of one electrode will be shifted eV relatively to the energy

levels of the other electrode, and a current start to flow through the barrier. This means that electrons

in the filled states near the Fermi level at the ferromagnet with the higher energy levels (FM1) tunnel

trough the insulator barrier to the empty states of the lower energy level ferromagnet (FM2)

Tunnelling Magnetoresistance

The first hypothesis in the model used to explain the tunneling magnetoresistance is that the

spin is conserved during the tunneling.

Assuming that the spin is conserved a spin up and a spin down current can be discriminated,

because spin up electrons coming from one ferromagnet only will be accepted by the unfilled spin up

states of the other ferromagnet and the same to spin down electrons.

The second hypothesis is that contribution to the tunneling current is directly proportional to

the density of states in FM1 times the density of states in FM2 at the Fermi level.

This phenomenon is well described considering the conductance, G, through the barrier.

In ferromagnets materials there is an imbalance in the density of states of spin up (U) and spin

down (D) direction that creates the magnetic moment in these materials.

When the two electrodes have a parallel magnetization, and supposing, for instance, that in

FM1 the majority of electrons with energy near the Fermi level have spin up they will tunnel to the

unfilled spin up states of the FM2 in which are also in majority (since the electrodes have the same

magnetization); regarding the spin down electrons will travel from a minority band to a minority band

(the band, which have a lesser density of states). So the conductance for the parallel configuration can

be written as:

where Ui and Di are the spin Up and spin Down densities of state in the ferromagnet i (i=1,2).

Figure 2.16: Tunnelling phenomenon

17

In the case of an anti-parallel magnetization of the ferromagnets the electrons are tunneling

from a minority band for a majority one, and vice versa. Thus, the conductance between electrodes

with parallel magnetization is higher than in the anti-parallel configuration.

Then the conductance in anti-parallel configuration is:

The difference in density of states at the Fermi level is expressed by:

P is the tunneling spin polarization.

The conductance change is:

Knowing that the resistance is the inverse of the conductance (R = 1 /G) we get to the

resistance change relatively to its lowest value (this is the TMR definition):

TMR is usually presented as a percentage which implies to multiply the previous expression

by 100. It is possible see that TMR higher for higher polarized electrodes, so they should highly

polarized materials.

Since the resistance is the inverse of the conductance from is possible to state that

a MTJ has a low resistance state when the ferromagnets have their magnetization parallel to each

other, and a high resistance state when they are in the anti-parallel configuration.

18

2.4.3.1. MTJ output response – linear and square signals

The resistance variation between the maximum and minimum values in a MTJ can fallow two

different rules: the MTJ can have a square of a linear response to the magnetic field. Linear responses

are used in sensors type MTJ, and the square behaviour is useful in data storage industry in order to

fabricate memories.

To obtain the desired response there are two distinct “easy axis configurations” to the MTJ: a

parallel anisotropy and a crossed one.

Parallel Anisotropy

The pinned and the free layer have both the same easy axis orientation:

The energy terms of the free layer for this configuration are:

Zeeman term:

Crystalline anisotropy term:

Demagnetizing field of the free layer term:

Demagnetizing field of pinned layer term:

Neel energy term:

The total energy of the free layer is:

In the previous equation is the magnetization of the free layer,

is the demagnetizing

field of the free layer and is the demagnetizing field of the pinned layer, which is given by

where N the demagnetizing factor and depends on the geometrical shape; finally we have

.

Figure 2.17: Scheme of a MTJ composed by a pinned layer and a free layer with parallel anisotropies.

19

In order to this energy to have a minimal value it is required that:

and

So the derivatives are:

For

there are 3 possible solutions:

1-

2-

3-

Substituting the first solution in the second derivative:

As the resistance variation of an MTJ follows the law the solutions of the

free layer’s energy minimization can show the transport behavior of such device. Two different cases

are obtained (Figure 2.18).

Figure 2.18: MTJ transport behaviour to the parallel anisotropy. Left: square response; Right: linear response.

Concluding, in parallel anisotropy is possible to obtain hysteretic transfer curves if ,

or linear curves if

. Therefore, if is desired a linear response from sensors that use these

parallel orientation for the easy axis of the free and fixed layer a strong demagnetizing field is required,

which is achieved by ensuring that the sensors present a large aspect ratio.

20

Crossed Anisotropy

The easy axis of the free layer has a direction perpendicular to the orientation of the easy axis

of the free layer:

Figure 2.19: Scheme of a MTJ composed by a pinned layer and a free layer with crossed anisotropies.

The energy terms of the free layer for this configuration are the same as previous with

exception to the anisotropy energy term, which is now , and so the total energy of the free

layer is:

Doing the same analysis as before the solutions that minimize the energy are:

1-

2-

Substituting the first solution in the second derivative:

So in the case where the pinned and free layer have crossed anisotropies the sensor

response will be always linear.

21

Figure 2.20: MTJ transport behaviour to the crossed anisotropy: linear response.

22

23

3. MgO MTJ biosensors for immunomagnetic lateral-flow

detection – System and sensor design [17],[19]

3.1. Reader system design

The sensitive part of the system is a magnetoresistive sensor which will detect the fringe field

produced by the magnetic particles. In order to the sensor identify the field created by the band with

the particles it is necessary that the band passes close to the sensor. This sweeping movement can

be done using two different methods:

1- Is used a static MR sensor on chip and the test strip mounted on a plastic wheel is brought in

contact with the sensor surface as it passes through it.

2- The second configuration uses also a static MR sensor mounted in a tape head and the

immobilized particles are pressed and passes through the read head (like a magnetic tape

passing over a read head).

Figure 3.1: Scheme- Magnetoresistive lateral flow detection – wheel concept; Picture- Close view of

the real wheel device.

In the case of the wheel device it is an functional prototype developed in INESC-MN and is

currently at INA (Instituto de Nanociencia de Aragón – Nanocience Institute of Aragon).

24

The so-called magnetic particles used in this project don’t have a magnetic moment per si;

they are paramagnetic particles so their magnetic moment in the absence of an external magnetic field

is null, meaning that they do not create any field. In order to magnetize the nanoparticles the wheel

device has a set of electromagnetic coils and a permanent magnet: the coils magnetize the particles

horizontally while the magnet magnetize them vertically.

Electromagnetic

coils

Wheel

MR sensor

Figure 3.2: Wheel device prototype

25

3.2. Sensor design

For this project MTJ devices are used as sensors (not as memories) it is required a linear

response and a low noise level. These two factors have conditioned the MTJ samples design (Figure

3.3).

In the mask used the are 21 sensors series, each with 360 individual elements (24 columns

with 15 elements each), placed in a central position of the die sample; 2 contacts are at the sides so

the stripe can be sweep freely and without any constrains over MTJ’s (such as wires bonding the

contacts to the chip-carrier). The mask also presents test structures: the two single element MTJs on

the right lower corner allow the comparison between their RxA and TMR values and RxA and TMR of

the series.

In order get a linear response the free layer of the sensors have a large aspect ratio: each

individual element is a rectangular pillar with 2 by 30 µm2. Concerning the noise level the MTJs were

arranged in arrays, constituting therefore series of MTJs so a lower detectivity could be achieved

Figure 3.3: Mask used in this work: sensors at the center and contacts at the sides (Appendix A for greater detail).

26

Figure 3.5: 3D scheme of the MTJs array’s: 4 individual MTJs linked in serious.

3.2.1. Noise and detectivity implications

Noise is a physical phenomenon present in every electrical circuit. Being generated by every

components present in circuits, either passive or active, it manifests itself by voltage fluctuations

across components and fluctuations in the current that flows through them.

The MTJs sensors signals are characterized by a transfer curve of their resistance as function

of the field. To get the value of their resistance at some applied field a current is forced through the

MTJ and the voltage at its terminals is measured. Consequently, in the specific case of MTJs, the

voltage fluctuations have the consequence of adding an uncertainty to the output value and so to the

resistance value. Thereby, the minimal variation in the magnetic field which can be detected depends

directly on the noise level.

The detectivity, D ( , is defined as

and is the minimum magnetic field magnitude that can be detected. It is directly proportional to the

output noise of the sensor at a certain band width, SV ( , and ΔV/ΔH is the slope of the transfer

curve at linear region (is the sensor sensitivity). The TMR of MTJs decreases with the applied bias

voltage, so the voltages applied should be small in order to operate with the maximum TMR ratio (at

the maximum TMR ratio of one sample we get the minimum D value).

Figure 3.4: MTJ series: array of MTJs; each individual element has a 30x2µm2 area.

=

27

The square of the spectral noise voltage density of in each individual junction is:

In this last equation the first term represents the white nose and the second term the 1/f noise,

where is the electron charge, I the baising current, V the voltage between the sensor electrodes of

an individual junction, T the temperature, kB the Boltzman constant, α the Hooge like parameter, A the

junction’s area and f the is the frequency. The white noise incorporates the thermal noise, which

results from the random thermal motion of electrons, and the shot noise, which results from the current

flowing trough discontinuities in the circuit, and is given by the Hooge model like expression.

If we consider a device with N junctions in series, each junction with a resistance r, with a

driven current I equal for all elements we have Voltage = N R I. Therefore the square of spectral noise

voltage density of such a series is given by:

For the same configuration (an array of N individual junctions) the sensitivity is:

So the square of the detectivity is:

and in the limit V<<kBT, the detectivity is then:

This last equation shows that for an array of N tunnel junctions the value of D decreases with

, meaning that the minimum field which is possible to detect is decreasing with this factor. This is

so because the square of the noise spectral density increases with N while the sensitivity (not the

square of the sensitivity) increases with N.

28

500 1000 1500 2000 2500 3000displacement m

0.0005

0.0000

0.0005

0.0010

V V

Functionalized stripe

Nanoparticle

3.2.2. Application: theoretical preview

In order to predict the outcome of a measurement of the magnetic particles over an array of

magnetic tunnel junction was assumed that the particles can be approximated to a magnetic dipole.

Also it was assumed that the particles are perfectly spherical and that the dipole center is at the

geometrical center of the particle. The magnetic field created by a magnetic dipole at the position r

from the dipole center is as follows:

As already described the magnetization of the particles can be either vertical or horizontal

depending on the use of the electromagnetic coils or the permanent magnet existing in the reader

system prototype.

In this thesis the theoretical prediction was made only for particles magnetized perpendicularly

to the sensor plane (vertical magnetization – along the zz direction) since this is the configuration

being currently tested with the reader device at INA.

As the magnetoresistive sensors used (which are indeed the focus of this work) only sense the

transverse in-plane component of the field created by the particle, only this component is relevant,

which was arbitrarily identified like the xx component.

Considering a spherical particle of radius r and magnetic moment

the xx

component of the magnetic field created is given by:

For the calculus of the expected behaviour of an array of MTJs was considered that the

magnetic particles had a susceptibility of 4.81, a 125 nm radius, and that they are placed at the

margins of the functionalized stripe. It was also assumed that the magnetic field at particles position

was a 5000Oe vertically oriented field (the correct technical data was not available). It was considered

a sensor series with 148 µm wide, and a pulling movement of a 740 µm wide stripe over the sensors.

The xx coordinate of the initial point was 1184 µm distant from the sensor. Due to the limitations of the

available computer and the highly time/processor consuming program only some points were

calculated.

Figure 3.6: Expected output: particles magnetized vertically (z coordinate =1000Å).

(µm)

V (V)

29

4. Micro fabrication – Systems, techniques and

processes protocols [13],[15],[16]

4.1. Sputtering Deposition Systems: Magnetron

Sputtering is a physical deposition process, whose foundations are based on the momentum

transfer between ions in a plasma a given target surface, against which they are accelerated. One way

of being able of deposit films by sputtering is using a magnetron system.

Figure 4.1: Scheme of a magnetron system.

The sputtering happens in a vacuum chamber, which contains the magnetron with the target

of the material to deposit and the substrate where it is intended to deposit. In the chamber is

introduced a gas (in a controlled way), which is subject to an electric field. This field is created by

applying a negative voltage to the target, which accelerates ions, always present in the gas despite

their overall neutrality. Depending on your energy, these ions are able to ionize neutral atoms by

colliding with the. These just-ionized atoms, are now charged particles, and so they starts to feel the

electric field originated by the target loads, thus they accelerates towards it, and participate also in the

ionization of new atoms, in what is a cascade process of successive ionizations. In the mean time the

electrons released in each ionization event also participate in this process, while having the

energy/momentum needed. These electrons would be repelled by the target; however thanks to the

presence of permanent magnets positioned behind the target, the electrons are confined to

trajectories in the vicinity of the target, along the field lines of the magnetic field. Thus, the process

described above, through the collisions of the considered species (electrons and positive ions), will

possibly results in the ignition of the plasma near the target. The magnetic confinement allows this to

happen with a lower concentration of atoms and therefore the pressure lower; however the use of

permanent magnets has the disadvantage of causing an irregular wear on target, revealing more

pronounced in the area where the ionization rate is higher, which corresponds to the region where the

magnetic field lines are parallel to the target surface (Figure 4.2)

30

Figure 4.2: Relation between the wear of a target and the magnetic field lines.

The ions that reach the target have energies/momentums in a wide range, depending on the

history of its route (when the ionization happened and the collisions they suffered), so the energy

available to be transferred to the target also belongs to a quite large range. Consequently the result of

a collision with the target is variable and directly dependent on the energy of ions. To lower energies

the impact of ions results in a simple perturbation of the structure of the target, resulting in an atomic

rearrangement that generates heat, which dissipates in the water cooling system existing on

magnetrons. In an intermediate energy range, the impact can cause the ionization of target atoms,

causing the ejection of electrons, captured by the magnetic field and therefore involved in the process

of ionization of plasma. Finally, for energies equal or greater than the bidding energy of target’s atoms,

the collision results in the emission of several of these atoms. Once they are neutral particles, the

target atoms travel freely over the magnetron chamber, and they are deposited on the substrate

forming a film that was intended to deposit.

This type of systems presents a major handicap concerning atomic species responsible by the

target sputtering: ions are charged particles, so that if the target is not conductor, the charges will

accumulate at the target, so after a while the field generated by them will shield the target and even

repel the other ion, stopping the sputtering process; this problem does not arise in case the target is a

conductor, because the ion charge is neutralized by electrons from the source of power. The solution

to this accumulation is doing the target biasing with an RF source instead of applying a DC voltage:

on the negative phase of the signal ions are attracted to the target, when it happens sputtering and

positive phase those that have accumulated on the target are neutralized.

The substrate, usually grounded, may also be applied an RF voltage. In this case, the

substrate becomes a target with which ions impact, thus making it possible the etching of the

substrate (sputter etch), which is useful both in the definition of structures at micro as in control

roughness of the deposited films.

31

4.1.1. Nordiko 2000

The Nordiko 2000 is an automated deposition sputtering system constituted by a deposition

chamber and a load-lock, which allows samples to be loaded and unloaded without breaking the

vacuum inside the main chamber (typical system pressure: 7x10-8

Torr).

The deposition chamber has six magnetrons positioned on top of the substrate table, with 3

inch diameter each. Between the substrate and the magnetrons is the shutter, a rotating disk with a

slot that can be placed beneath the targets. The substrate table has 12 slots for substrate holders,

spread by four quadrants of 3 slots each. These, three are water-cooled and electrically grounded,

and one is heated and electrically isolated from the rest of the table so that can be connected to

ground or to a power supply allowing surface treatment by sputter etch. One slot in one of the water-

cooled quadrants has an array of magnets, providing a 30 Oe magnetic field for easy axis definition

when magnetic materials are deposited.

Target # Material Magnetron parameters Ar Flow (sccm) Pressure (mTorr) Dep. rate (Å/s)

1 Mn60%Pt40% DC:40mA 9 5.0 0.64

2 Ru DC:45mA 8 5.0 0.44

3 Ta DC:45mA 10 4.5 0.52

4 Co40%Fe40%B20% DC:45mA 9 5.0 0.44

5 MgO RF:130W 9 18.0 0.142

6 CoFe RF:35W 8 5.0 0.36

Table 4.1: Deposition setpoints of Nordiko 2000.

Figure 4.3: Nordiko 2000 deposition camber schematic.

32

4.1.2. Nordiko 7000

Nordiko 7000, like Nordiko 2000, is also a fully automated commercial magnetron. Being able

of handle 6 inch industrials wafers, consist in a central dealer camber, connected to four process

modules, and a load-lock (Figure 4.4).

Figure 4.4: Nordiko 7000 - disposition of the all process modules, dealer and load-lock.

In the load-lock, pumped by a turbo pump, is achieved a “primary” vacuum, with a base

pressure about 5.0x10-6

Torr. The dealer and each process chamber are pumped with cryogenic

pumps, reaching base pressures of the order of 5x10-9

Torr. A robot arm, placed inside the dealer,

carries the wafers from the load-lock to each process module (Figure 4.4).

As can be seen in the previous figure, each module has a different purpose.

Module 1 – Used to perform flash annealing processes (wasn’t used in the present work).

Module 2 – Perform a soft sputter etch, usually used to remove small a thickness of natural oxidized

material, before the deposition of TiW(N) and Al; a argon (Ar) plasma and two RF power sources are

used, one applied directly to the target and primarily responsible by speed up the ions towards the

target (RF1 source), and another source exclusively responsible for the plasma maintenance by

promoting his ionization.

Module 3 – Carries out a reactive sputtering deposition of TiW(N). During a reactive sputtering at least

one gas is introduced in the process chamber with a purpose beyond the maintenance of the plasma.

In this case, the deposition is made from a TiW target with a mixed Ar/N2 plasma, wich provides the N2

required to react with the target material and become incorporated in the deposited film. This material

was used has passivation layer because:

- Is a hard material, protecting the sample against physical damaged;

- It is dense, offering a good chemical protection against corrosion caused by substances used

in sample fabrication (like the revelator used after the lithography);

- Is a dark metal, so it’s used as anti-reflective coating for the direct opto-lithography, minimizing

reflected light.

33

Module 4 – Used to the deposition of Al98.5Si1.0Cu0.5. It’s used an Ar plasma powered by a DC power

supply.

The plasma is maintained with a DC power supply, and the standard deposition conditions can

be seen in the table above.

Operation RF1 RF2 DC Pressure Gas Flow

Module2 Soft Sputter

Etch

70W 40W - 3.0mTorr 50 sccm Ar

Module3 TiW(N) - - 0.50kW 3.0mTorr 50 sccm Ar+10 sccm N2

Module4 Al - - 2.0kW 3.0mTorr 50 sccm Ar

Table 4.2: Standard conditions operation in the used modules

4.1.3. Ultra High Vacuum-UHVII – Oxide Sputtering System

Another sputtering deposition machine, UHVII is used to deposit Al2O3. This system, a 6 inch

magnetron, has a single chamber, so there is no load-lock, and so it isn’t possible to load or unload

samples without vent the process chamber; this has a consequence materialized in the time need to

reach the optimal 3x10-7

base pressure for deposition: roughly 12h hours. An Ar plasma powered by a

RF source sputters a ceramic Al2O3 target placed above a 6 inch static and water-cooled substrate

table. Obviously 6 inch wafers can be used in UHVII, as well as smaller area samples; trough this 6

inch usable area the uniformity of the deposition varies about 10% between the center and the edge of

the table.

Figure 4.5: Schematic side view of UHVII.

34

Figure 4.6: Picture of UHVII.

4.2. Ion Beam Systems

A Ion Beam Deposition (IBD) systems can be used for deposition of thin films and for ion

beam milling, which consists in a non selective dry etch process. In IBD systems, a highly energetic

ion beam is used to remove material from a target that will deposit on a substrate a thin film in a

deposition process, or used to remove material from the sample itself in the etching process. The ion

beam is created by an ion source, called ion gun, within which is created a plasma by an RF power

supply. The ions are extracted from the gun by a set of three charged grids, which pull out the ions

from the plasma and accelerate them towards the vacuum chamber as a uniform collimated beam that

will impact a target.

There are two ion guns in the IBD system: the deposition gun and the assist gun. The

deposition gun points to the targets made of the material to be deposited. In the deposition process

the ions that hit the target are less energetic than in the sputter process, implying a lower deposition

rate.

In the assist gun, the beam points directly to the substrate and can be used to ion milling

process. Besides assist and deposition guns, there are two neutralizer guns inside the chamber, which

emit electrons to neutralize the accumulated ions on the target and on the insulating substrates .The

assist and deposition guns, the substrate and the target are disposed in a Z configuration.

Material RF Pressure Gas Flow Deposition Rate (Å/s)

Al2O3 200 W 3 mTorr 45 sccm Ar ~0.1905 at the center ~0.1714 at the edge

Table 4.3: Deposition conditions of UHV2.

35

Figure 4.7: Schematic view of a IBD system: Z configuration. The substrate table has a permanent magnet array mounted around it, producing a 40 Oe

magnetic field that defines the easy axis during the deposition. This table can be rotated depending on

whether a deposition or an etch is being made. The substrate holder also rotates during deposition

and milling processes in order to achieve a better uniformity throughout the sample. The target holder

has a hexagonal prism shape with one different target in each face, and rotates according to the

material that will be deposited, letting a specific target exposed to the ion beam, while the others

targets remain protect from contamination by a shutter.

There is also another shutter protecting the sample until assist and deposition guns have the

ion beams stable, accordingly to the set parameters. The chamber vacuum is obtained with a turbo-

molecular pump and a cryogenic pump, achieving a working base pressure of 10-8

Torr.

There are two IBD systems in INESC, the Nordiko 3600 and the Nordiko 3000, both installed in a

class 100 clean-room. They are both very similar to each other, where the main difference is that

N3600 is a much larger system being capable to handle 8 inch substrates, instead of the 6 inch

provided by N3000. During this work both facilities were only used as ion milling tools.

System operation RF V+ V- Gas Flow I

N 3000 Ion milling 58W 500V 200V 8 sccm Ar 30 mA

N 3600 Ion milling 160W 735V 350V 10 sccm Ar 105 mA

Table 4.4: Ion milling used set values for assist gun of both IBD systems.

36

4.3. Pattern Transfer - Techniques and Processes

For the fabrication of devices, and considering they are allowed with sizes ranging from

microns, it’s fundamental the ability to selectively remove or deposit material from a substrate. In this

work were two the techniques used to do so, namely lift-off and etching techniques.

4.3.1. Pattern Transfer Techniques

Etching

This process starts with a substrate in which was deposited the material to be patterned. That

material is selectively covered with a mask, so, this way, fractions of the sample become protected.

Then the not protected material is removed, and at the end the mask is itself removed, leaving the

substrate with the desired shape.

1) 2)

3) 4)

Lift-off

As well as in etching, also in lift-off process the ultimate goal is to obtain a material with a

certain pattern. Again, it still is used a mask but it is laid on a substrate without the material intended to

shape; that material is deposited over the substrate with the mask already on it, and then the mask is

removed taking with it all the material covering it, resulting in a substrate with a patterning material and

free material areas where previously was laid the mask.

1) 2)

3) 4)

Figure 4.8: Etching process.

Figure 4.9: Lift-off process

37

4.3.2. Direct Write Laser Optical Lithography (DWL)

In order of being able perform any of the two techniques described above a mask must be

transferred to the substrate. There are two types of masks that can be used, hard masks and soft

masks, and several methods to create and apply them. In this work all the masks were created using

direct write laser optical lithography process which includes mask design, vapour prime, photo-resist

coating, lithography exposure and photo-resist development. The masks were designed in a CAD

software, converted in a set of binary files and transferred to the lithography system hard drive.

4.3.2.1. Vapour prime system

This is a system placed in the clean-room that helps to ensure that the optical lithography is

successfully done. Basically the vapour prime system is an oven, pumped to remove any unwanted

material and heated to prevent any residual water molecules of remaining in the sample surface.

When the oven is completely pumped down a gas is introduced in his chamber: hexamethyldisilizane

(HDMS), is inserted to promote a better adhesion of the photo-resist onto the sample; after 5 minutes

in this hexamethyldisilizane atmosphere approximately one monolayer of HDMS will be coated at the

sample surface. The steps that compose the process which the sample undergoes during the vapour

prime are described in detail below (program 0).

1. Wafer dehydratation and purge oxygen from the chamber. Vacuum, 10 Torr, 2 minutes.

N2 inlet, 760 Torr, 3 minutes x 3

times.

2. Priming: Vacuum, 1 Torr, 3 minutes.

HDMS, 6 Torr, 5 minutes.

3. Purge prime exhaust: Vacuum, 4 Torr, 1 minute.

N2 inlet, 500 Torr, 2 minutes.

Vacuum, 4 Torr, 2 minutes.

4. Return to atmosphere pressure: N2 inlet, 3 minutes.

38

4.3.2.2. Silicon Valley Group Track – Coating and Developing Track

After the vapour prime, the samples are coated with the photo-resist, which is done in the

Silicon Valley Group Track. Being a fully automated system the track performs two main tasks:

deposition of 1.5 µm thick positive photo-resist (PFR 7790G 27cP) and its development after it had

been exposed. This photoresist is made of a resin and a photo-reactive compound dissolved in a

solvent that is also dissolution inhibitor. Because of the photo-reactive compound the photo-resist is

light sensitive, becoming unstable under exposure to certain light wavelength. The thickness of the

photo-resist is defined by the angular velocity in the spinning module in which the samples are coated

with photo-resist: for 40 seconds the spinning module makes 2800 revolutions per minute.

Figure 4.10: Picture of the SVG tracks.

Conditions used for the coating and developing processes:

Coating – Recipe 6/2

- Dispense of photoresist at 800 rpm for 5 seconds.

- Spin at 2.5 krpm for 40 seconds, then at 1.6 krpm for 5 seconds.

- Clean photoresist form the wafer border at 1 krpm for 2 seconds.

- Spin at 1.5 krpm for 15 seconds.

- Bake at 100ºC for 60 seconds.

Developing – Recipe 6/2

- Bake at 110ºC for 60 seconds.

- Cooling for 30 seconds.

- Water spray rinse at 500 rpm for 1 second.

- Dispense of the developer (Ethyl lactase – Pth70eg) at 500 rpm for 5 seconds.

- Development for 60 seconds with the wafer stopped.

- Rinse with DI water at 1 krpm for 20 seconds.

- Drying with wafer rotating at 3.5 krpm for 30 seconds.

39

4.3.2.3. Optical Lithography Exposure

The optical lithography exposure is performed by a direct-writ laser system DWL 2.0

Lasararay, which makes use of a 442 nm Helium-Cadmium laser of 120 mW. The laser sweeps the

sample, turning on and off concordantly with a predefined mask designed by the user with proper CAD

software (AutoCAD). The write lens is focused on the sample surface by an air pressure auto-focus

system, and exposure energy can be adjusted accordingly with reflectivity of the substrate material.

Since the used photo-resist is a positive one, it became unstable in the areas where the laser was on.

Following the sample exposure a photo-resist development is necessary: in this step the unstable

parts of the photo-resist vanish in contact with the developer(Ethyl lactate-commercial name Pth70eg).

1) 2)

3) 4)

For the exposure the samples are mounted on a mechanical x – y stage and fixed by vacuum.

In samples alignment is used a 70 nm measurement accuracy dual CCD camera system

(macro/micro). During the exposure the laser sweeps the sample according to the designed mask,

scanning the sample in stripes 200µm wide. Each stripe corresponds to several scans performed by

successively writing pixels from left to right (pixel grid pitch: 200 nm), in the x direction. Finished one

scan, the sample is moved one step in the y-direction and the next scan is done. When the stripe

exposure is complete, the stage moves to the origin in the y-direction and 200µm to the right in the x-

direction, starting the next stripe.

The system existing in INESC’s is placed in a class 10 clean-room and is capable of defining

structures of about 1 µm. The masks exposed were previously converted to the DWL software and

load into the hard disk of the system. Their design is done in the world spread AutoCAD software.

Figure 4.12: Figure of the DWL system.

Figure 4.11: Optical lithography – photo-resist mask creation.

40

4.4. Annealing setup

After completely patterned, it still is needed a magnetic thermal treatment to each MTJ sample

get fully functional. This annealing promotes the crystallization of amorphous materials, and the re-

crystallization of already crystalline structure correcting the defects always existing in their as-

deposited crystalline structure, re-enforcing the magnetic layers magnetization and improving the

exchange magnetic field that fixes the pinned layer. Once the annealing temperature for MTJs is

higher than the blocking temperature of the antiferromagnetic (pinning layer) is necessary to apply a

magnetic field during the annealing which re-sets the exchange field with the pinned layer. There are

two annealing setups at INESC-MN: it was used the system showed below.

Figure 4.13: Picture of the used annealing setup.

The setup consists in a heating system, a removable glass chamber and two large water-

cooled coils around a movable soft iron core. The heat source is a halogen lamp (100W, 12V) inside a

cooper block. The samples are placed on top of the cooper block with the glass chamber positioned

around and sealed, involving samples and heating system. The chamber is pumped by a turbo pump,

reaching pressures of the order of 10-6

Torr, and a magnetic field is applied by the water-cooled

electromagnetic coils. Under the magnetic field, the samples are then heated up to a certain

temperature defined by the user, and then still under the magnetic field, left to cool naturally. This is

the main difference between the two annealing setups: unlike this one this one in the other setup the

samples are heat up without an applied field.

All the samples patterned in the course of this thesis were annealed at 320ºC for 1 hour under

a magnetic field of 4kOe. The temperature cycle, which the samples are submitted to, is as follows:

41

Figure 4.14: Temperature cycle in the annealing process.

42

43

5. Characterization methods – Measurement techniques

[13],[14],[15],[16]

5.1. Vibrating Sample Magnetometer (VSM)

The VSM system (Figure 38) measures the magnetic moment of some material. It allows the

magnetic moment measurement of bulk unpatterned samples as a function of the applied magnetic

field. Because is a characterization method that doesn’t require pattern samples it can be precious by

saving time that would be wasted processing samples with an inappropriate magnetic response.

Two large coils are responsible for the creation of the magnetic field which defines the sample

magnetization. That is mounted onto a quartz rod which is connected to piezoelectric crystal, which

under excitation makes the sample vibrate. In the region between the large coils, near the sample, two

smaller coils were placed. Being a magnetic material the sample creates a magnetic field collected by

the small coils. Because the sample is moving back and forward the magnetic flux crossing the plane

of the inner coils is not constant, and a current, proportional to the variation rate of the flux, will be

induced on them. That current depends naturally on the magnetic moment of the sample.

Figure 5.1: VSM system.

5.2. Manual Transport Measurement Setup

This system used for electrical characterization of patterned samples. It comprises two pairs of

micropositioning probes with TiW needles (10 µm resolution), one voltmeter, two current sources (one

for the sensors biasing and other to create the applied magnetic field), 2 Helmoltz coils and magnifying

lenses.

Transfer curves (curve of Resistance vs external DC magnetic field) were measured in this

setup: curves of device resistance vs applied field. A current source is connected to a top and a

-6.0E-04

-3.0E-04

0.0E+00

3.0E-04

6.0E-04

-6.0E+03 -4.0E+03 -2.0E+03 0.0E+00 2.0E+03 4.0E+03 6.0E+03

M (

e.m

.u)

H (Oe)

VSM unpatterned

sample

T annealing = 320 ˚C

Hex = 1300 Oe

44

bottom contact of a sample, as well as the voltmeter (also connected to top and bottom contacts). It is

through the needles that electrical contact is established. The magnetic field is created by two coils

powered by the second current source, and varies sweeps all the range between -140 and 140 Oe. All

this sources are controlled by a computer with proper software (which also records the data) and a

GPIB communication protocol

The measurements can be done with two or four probes, being preferable the four probes way

whenever is possible.

It is considered that the real voltmeter behaves close to an ideal one: if it hasn’t infinite

impedance however, still should have a very high one, so current passes through him.

With the two probes configuration we have,

and the TMR comes:

In the last equation it’s clear that with two probes the measured TMR is smaller than actual

TMR value.

Because no current flows through the voltmeter, a four probe arrangement allows do

not count with contact resistance in the TMR measurement.

In the present work all the series were measured with a two probes arrangement, which,

because of the series design itself was impossible to measure with four probes.

Figure 5.2: Electrical scheme for measurements with 2 (on the left) and four (the right one) probes.

45

5.3. Profilometer

The prodfilometer system allows to measure the topography of a sample’s surface through a

piezo-resistive sensor. This is a direct measure where a micro-needle (12.5 µm diameter) makes a

physical contact with the surface to study, and sweeps it in a defined range. This system is very useful

to measure thicknesses of films and oxides. In the work contained in this thesis the profilometer was

mainly used to validate the thickness of deposited oxides.

To be able to measure the thickness of deposited material is required the existence of a

boundary between a region that contains the sample film/oxide and another region where it is not

deposited (the latter is a region where the substrate is uncovered at the time of profilometer measure).

There will therefore, in these transition regions a gap that corresponds to the thickness of the film. To

get these “undeposited” areas the substrate is, before the deposition, partially covered with some

material that can be easily removed afterwards (and with it, all the material deposited in top of it will be

removed too).

As the needle travels to the sample surface, the profilometer traces a graph with the height of

the needle. The value for height of the needle is given in each moment for a given reference level.

Usually the profilometer was used in parallel with the ellipsometer to ensure the right

measurement of the oxide thickness.

Substrate

Deposited material Deposited material

Figure 5.3: INESC-MN profilometer system – Allows measure film topography with irregularities higher then 300Å, with a 5Å vertical resolution

Figure 5.4:

46

5.4. Ellipsometer

The AutoEl ellipsometer, like the profilometer was mainly used to measure the thickness of

calibration samples used during the deposition of Al2O3. The principle of operation is based on the

irradiation of the sample’s surface with a collimated beam of monochromatic light (wave length

λ=632.8 nm) from a known angle. The incident light has a controlled state of polarization, and the

difference between this state of polarization and the one of the reflected lights is determined and

quantified in terms of two angles, Δ and Ψ, which are function of the real and imaginary components

of the refraction index. Then, using a numerical algorithm, the ellipsometer returns the refraction index

and the thickness of the layer. It’s important to be aware that this model assumes a perfectly reflective

substrate: this mean that every sample measured in the ellipsometer is grown on top of a Si substrate.

Figure 5.5: Picture of the ellipsometer.

47

6. Microfabrication – Process steps

6.1. Stack deposition

The first step of every new process run is the deposition of the sample stack over a new, clean

substrate. In this work the different material layers that compose a sensor were deposited in the

already presented Nordiko 2000 machine. Afterwards is necessary deposit a passivation layer of

TiWN2 to avoid the oxidation of the top layer, to add some physical protection to the sample stack and

to act as a antireflective layer to the optical lithography. Before the TiWN2 deposition a soft etch is

performed to remove a small thickness of natural oxidized material. This deposition took place in

Nordiko 7000 system.

Figure 6.1: Left - Standard MTJ stack used; Right - 3D model of a bulk MTJ sample after the stack deposition

Ta 50Å

Ru 50Å

CoFeB YÅ

MgO XÅ

CoFeB 30Å

Ru 9Å

CoFe 20Å

MnPt 200Å

Ta 30Å

Ru 180Å

Ta 50Å

Substrate

(Si or glass)

Substrate

MTJ stack

Oxide barrierMgO barrier

MTJ stack deposition

48

6.2. Bottom contact definition

To define the bottom contact of the sensors a lithography, followed by an all stack etching, are

realized. However, before the exposure there are some procedures that must always be met: first of

all, the samples are put in the Vapour Prime oven for 30 minutes at 130ºC, while HDMS is introduced

in the chamber; then the samples are coated in the SVG coating track with a 1.5 µm thick photoresist.

After the coating the samples are exposed and the laser makes an impression of the mask onto the

resist.

Following the lithography the samples are developed in the SVG developing track, in wich the

developmente liquid (usually called developer) is poured onto the sample, with which stays in contact

for 30 seconds. At the end of that time the sample is spread with water to wash the developer,

preventing overdevelopment.

After an optical inspection to confirm that the exposure was coorectly done the samples are

ready to be etched.

The next step is the first ion milling to define the bottom contact, performed either in N3000 or

N3600 system. This is a non selective etch, purely physical phenomenon (linear moment transfer),

both photoresist and stack are etched, but because the photoresist is much thicker (1.5 µm) than the

MTJ structure (around 800 Å) it never be fully etched.

Figure 6.3: Bottom contact definition: resulting structure after the ion milling.

Figure 6.2: Resulting photoresist after the 1st lithography revelation.

1.5 nm

Substrate

MTJ stack

Photoresist -

Photoresist1.5 µm

8 µm

Substrate

1.5 µmIon milling70˚ pan

8 µm

49

Figure 6.4: Picture of bottom contact.

Once the etch is donne th resist strip is required in order to remove the remaining photoresist.

This is donne immersing the samples into a microstrip soluction, which acts like a photoresist solvent.

An ultrassonic mach is also used to help remove the photoresist.

Figure 6.5: Left - Wet bench’s picture: ultrasound machine and hot bath; Right - microstrip solution.

148 µm

Hot bath

Ultrasound

machine

50

Substrate

2 µm

30 µm

6.3. MTJ pillar definition

The bottom contact is defined and is now necessary to define the junction pillar. This is done

again with a lithography followed by an etching. All the steps preceding any lithography have to be

fulfilled.

This second lithography is very important because it has to be well aligned with the previous

defined layer (bottom contact layer). The alignment between two different layers is always done with

the help of alignment marks drown in previously defined layers. Also the development is very

important because of the aspect ratio, which is essential to get a correct response of the sensors, so

an overdevelopment has to be avoided.

The next etching is also crucial because there is a certain thickness of material that has to be

removed, but not all the film layers; so it is vital prevent over-etch: the etching has to be performed till

under the SAF, and stop before reach the substrate (preferable is a good practice stop the etch some

hundreds before the end of the stack). To control the ending point are used calibration sample, which

have only and only a structure corresponding to the part of the MTJ stack that it is supposed to be

removed (it can have more 50Å of material above the SAF to ensure that is level is passed during the

etch). Another precaution to have during this ion milling regards the angle between the sample and the

ion beam: although using a 70º minimizes shadow effect, it implies an important effect of redeposition,

which can short circuit the oxide barrier, so a 40º angle is used to minimize redeposition.

Figure 6.7: Picture of the pillar junction.

Figure 6.6: Pillar junction definition. Left - After the 2nd

lithography; Right – After the 2nd

etch

Substrate

Bottom contact

MTJ pillar

Ion milling40º pan

51

Substrate

Bottom contact

Oxide800Å Al2O3

6.4. Pillar lateral isolation

In the samples microfabrication sequence fallows an oxide deposition. An 800Å of Al2O3 layer

is deposited to isolate the barrier so that the electrons only flow through the junction, avoiding all the

short circuits that would happen at the top contact deposition if the oxide had not been deposited.

After the oxide thickness confirmation in the profilometer and/or in the ellipsometer, the coming step is

the oxide lift off. The samples are kept in hot bath (65ºC) immersed in a microstrip solution used like in

every lift off, and the ultrasound machine is used as often as possible till the lift off become complete.

6.5. Top contact definition – Metallization

A third exposure defines the mask that will define the top contacts. This is exposure is inverted

when compared with the all the orders: this means that, now, after the development is intended to do

not have photoresist in the areas of interest, which are the places where the metal is intended to be

deposited (unlike the previous lithography in which the areas of interest remained covered and

protected).

The metallization is performed in the N7000 system, where 1500 Å of aluminum are deposited

after a prior soft etch. On top of the aluminum a 150 Å layer of TiWN2 is deposited to protect the

sample to the last lithography. A lift off removes the metal outside the contacts.

Figure 6.8: MTJ pillar and bottom contact insulation: after the lift off.

Figure 6.9: Top contact definition.

Substrate

Bottom contact

Top Contact1000Å Al + 150Å TiWN2

52

6.6. Sample insulation

A final lithography is made to define some holes in the contacts. After the development some

small photoresist blocks are on top of the contacts. Then a 1000 Å Al2O3 oxide layer is deposited,

followed by one last oxide lift off which opens small holes in the oxide through which the electrical

contact can be done with the sample.

6.7. Sample dicing

Since that in each substrate more than one die is patterned it is necessary to “cute” the

sample, separating each die and make it an individual module. The dicing is done by an automatic

dicing saw, model Disco DAD312, which can process up to 6 inch diameter substrate and 30.6 mm

thick.

6.8. Annealing

Finally, the samples need a magnetic thermal treatment: they need to be annealed at 320˚C

under a 4 kOe magnetic field during 1 hour, in order to promote the re-crystallization of the as-deposit

materials, re-enforcing the magnetization of the magnetic ones, and improving the TMR of the

samples.

6.9. Encapsulation

The last process step is mount the individual die on a chip-carrier (chip-mounting) and connect

them each other (electrically connect the sample and the chip carrier with very thin low resistance

wires – Wire bonding).

148

µm

Figure 6.10: Picture of the top electrode.

Figure 6.11: Sample patterned and mounted on a chip-carrier

53

6.00E+04

8.00E+04

1.00E+05

1.20E+05

1.40E+05

-150 -110 -70 -30 10 50 90 130

R (

Oh

m)

H (Oe)

ZarMTJ1

TMR = 78.1%Sensibility = 0.426%/OeVbias =1.7601 V

7. Results

During this thesis two generations of samples were patterned: the first generation was

patterned in a Si substrate coated with a 800Å thick SiO3 layer, and the second generation used glass

substrates. For both generations several samples were processed with distinct stacks: different

thicknesses of MgO barrier and CoFeB free layer weere tested.

1st Generation – SiO2 substrate

The used MTJ series have 4 contacts, two at the extremities and two intermediate contacts

which should allow a study in function of the number of individual MTJ.

Each of these three sections (sections ab, bc and cd) of the MTJs array has 120 individual

elements. The graphics below represent measurements of the entire series, all the 360 elements.

ZarMTJ1 ZarMTJ2 ZarMTJ3 ZarMTJ4 ZarMTJ5

Mgo thickness (Å) 17 17 12 12 12

CoFeB thickness (Å) 100 15,5 15,5 30 60

Table 7.1: Variations in the stack of samples with Si substrates.

Figure 7.1: Close view of one MTJ series with the 4 contacts identified as a, b, c and d.

Figure 7.2: ZarMTJ1 transfer curve

54

Figure 7.5: ZarMTJ5 transfer curve.

1.10E+04

1.15E+04

1.20E+04

1.25E+04

1.30E+04

1.35E+04

-150 -100 -50 0 50 100 150

R(O

hm

)

H(Oe)

ZarMTJ2

TMR = 17,04%Sensibility = 0.164%/OeVbias = 1.7173 V

Figure 7.4: ZarMTJ3 transfer curve.

This previous graphics represents the typical transfer curve that is possible to obtain from

each of the 1st generation samples (there are only four graphics because the sample ZarMTJ4 didn’t

make through the patterning process). These TMR values are not great for any of the samples, since,

at least, values around 100% were expected. Looking at samples ZarMTJ1 and ZarMTJ3 graphics we

would say that at least the relative sample behaviour was right, since from the thicker free layer

ZarMTJ1 sample we would expect a higher TMR comparing with the thinner free layer ZarMTJ3,

which should have in turn a higher sensitivity. In fact, these two samples behave just like that when

compared to each other. However the other two graphics show some problems: the one from

ZarMTJ2 sample shows a smaller sensibitility than the sample number 1 even though it had a thinner

free layer; the ZarMTJ5 has by far the worst TMR (it was not expected) and the highest coercivity of

the set.

Figure 7.3: ZarMTJ2 transfer curve.

55

In the graphic below the previous curves of ZarMTJ1 and ZarMTJ3 are overlapped for a

better comparison of their sensibilities.

Figure 7.6: ZarMTJ1 and ZarMTJ3; measurement for 360 elements.

The previous graphic evidences the relative correct behavior of these two samples: as the free

layer thickness decreases the sensibility should increase.

Figure 7.7: The RxA of the sensors in ZarMTJ1 and ZarMTJ2 samples. Each sample has 21 independent series of MTJ, here identified by a number line.

0

0.2

0.4

0.6

0.8

1

-150 -100 -50 0 50 100 150

No

rmal

ize

d T

MR

(%

)

H (Oe)

ZarMTJ3

ZarMTJ1

MgO– 12ÅCoFeB– 15.5ÅTMR = 23.6 %Sensibility = 1.241%/Oe

MgO– 17ÅCoFeB– 100ÅTMR = 78.1%Sensibility= 0.426 %/Oe

56

The RxA of the first two generations (ZarMTJ1 and ZarMTJ2) should be similar since they

have the same MgO thickness. However it is clearly shown in Figure 7.7 that the RxA of the ZarMTJ2

is much smaller. Since these samples were deposited and processed at different moments (not at the

same time or with very few time between them) this means that the reproducibility of the entire

process starting in the deposition (inclusive) through all the patterning protocols until the

measurements it is not reliable. Therefore comparisons between the samples, in order to define a

better stack cannot be done (even the ones already done only would be entirely correct if all the

samples had been processed at the same time).

Although this non reproducibility problem, the study of the behavior of a series based on the

number of elements that compose it would still be possible.

Figure 7.8: ZarMTJ1 - TMR vs number of junctions.

Since it isn’t reasonable to expect that all the junctions in a series are equal, the TMR and

resistance values of one series should be a weighted average of their individual values. With this in

mind, at the first sight the previous figure seems quite nice: a majority of series have more or less the

same TMR value for 120, 240 or 360 individual elements, meaning that they are quite similar through

the array lenght. But complementing this graph with some extra data raised some questions. Looking

at the series number 6:

TMR(%) Rmax (kΩ)

line ab(#120) bc(#120) cd(#120) ab(#120) bc(#120) cd(#120)

6 63.95 51.09 51.11 68.20 49.30 41.10

With these values in the sections ab, bc and cd, when the section ad is measured the TMR

can’t be lower than 51.11% and the resistance should be 158.6 kΩ. Nonetheless the measured

resistance is 92.10 kΩ with 44.62% of TMR (table above).

0

10

20

30

40

50

60

70

80

90

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

TMR

(%

)

line

360 junctions

240 junctions

120 junctions

ZarMTJ1:100Å CoFeB17Å MgOIbias = 20 µA

57

TMR(%) measured Rmax "expected" (kΩ) Rmax measured (kΩ)

line ac(#240) ad(#360) ac(#240) ad(#360) ac(#240) ad(#360)

6 58.9 44.62 117.50 158.60 117.00 92.10

This non expected kind of behaviour can be verified in other series, in other dies and even in

other samples. Because of the decreasing of the resistance relatively to what was expected be this

effect seems like the one obtained with resistors mounted in parallel. In fact this behaviour warned that

there might be some leak of current to the substrates. Subsequent measurements confirmed that

there were current leaks through the oxide that is supposed to insulate the Si substrate. Was then

determined that all the series were this way connected with each others. It was even possible to get

some transfer curves making the contact through pads that should be isolated one from the other.

Figure 7.9: Transfer curve between contacts of two independent series.

Every 1st generation samples had this current leak problem, which obviously affects the

performance of the magnetic tunnel junctions, which were deposited over a silica substrate with 800Å

of SiO2.

2st Generation – Glass substrate

The second sample generation was processed over a glass substrate and the conduction

problem disappeared. Meanwhile, was discovered by the team of INESC researchers that the

deposition rates in the Nordiko2000 system didn’t correspond to the known values. So in order to

regain the control of the thickness of deposited films the deposition times were recalibrated. This new

glass substrate samples were already deposited with the new deposition times.

5.00E+04

6.00E+04

7.00E+04

8.00E+04

9.00E+04

-180 -140 -100 -60 -20 20 60 100 140 180

R (

Oh

m)

H(Oe)

ZarMTJ1 line 12 pad a & line 13 pad dTMR = 36.7%

58

7500

7700

7900

8100

8300

8500

-150 -50 50 150

R (

kO

hm

)

H (Oe)

N2#4

TMR = 3.8%Sensibility = 0.033%/OeVbias = 1.82 V

4.00E+04

4.50E+04

5.00E+04

5.50E+04

6.00E+04

-150 -50 50 150

R (O

hm

)

H (Oe)

N2#1

TMR = 39.5%Sensibility = 0.711%/OeVbias = 1.79 V

16000

18000

20000

22000

24000

-150 -50 50 150

R (k

Oh

m)

H (Oe)

N2#2

TMR =32.0%Sensibility=0.1723%/OeVbias = 1.80 V

N2#1 N2#2 N2#3 N2#4

CoFeB thickness (Å) 15.5 30 100 60

MgO thickness (Å) 12 12 12 12

Table 7.2: Variations in the stack of samples with glass substrates.

Figure 7.10: N2#1 transfer curve.

Figure 7.11: N2#2 transfer curve.

Figure 7.12: N2#4 transfer curve.

59

All these samples were patterned at the same time, so they could be compared with each

other. The sample N2#3 (100 Å CoFeB) was not working at the end of the process. Comparing with

the 1st generation, the sample with 15.5Å of CoFeB showed an improved magnetoristance although

the decrease of the sensibility and the increase of the coercivity. For the sample N2#2 was expected

higher TMR values than N2#1 sample. The sample N2#4 has marginal TMR. The sample with higher

TMR, N2#1 sample, was chosen to further studies.

Once more a study of the behavior of a series based on the number of elements that

compose it was conducted. The results of the sample N2#1 are shown below:

Figure 7.13: N2#1 - TMR vs number of junctions

The TMR values for 120, 240 and 360 individual elements are almost the same, close to what

would be expected from perfect series concerning the likeness of every individual junction. Since

during the patterning process of both ZarMTJ1 and N2#1 nothing occurred that could explain a lesser

uniformity of properly working individual junctions through a series, again can be concluded that was

the conduction path through the substrate in the case of the first generation samples that played a

major influence in the unexpected behavior of ZarMTJ1 (and all the other 1st generation samples), and

not an unidentified cause in the manufacturing process.

0

10

20

30

40

50

0 5 10 15 20

TMR

(%)

Line

360 junctions

240 junctions

120 junctions

sample N2#1Ibias = 40 µA

60

0

10

20

30

40

50

60

70

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

TMR

(%)

V(V)

N2#1line 18 ab (#120)

A study of TMR vs Vbias was conducted (this study required a remake of the program that

controls the manual transport measurement setup).

0

10

20

30

40

50

60

70

-35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35

TMR

(%)

V(V)

N2#1line 18 ab (#120)

-35

-30

-25

-20

-15

-10

-5

0

5

10

15

20

25

30

35

-0.003 -0.002 -0.001 0 0.001 0.002 0.003V (V

)

I (A)

H = 140 Oe

H = -140 Oe

N2#1line 18 ab (#120)

Figure 7.14: Top – Curves V vs I for maximum and minimum resistance saturation fields; Center – TMR vs I; Bottom – Close look of the near zero bias current.

61

It can be seen that the TMR value has a plateau for currents in the vicinity of I=0 (V=0). In fact

it is possible to apply about 1.5V to the series without any concern about the decrease of TMR. If the

current keeps increasing the TMR starts to decrease. With the used voltmeter wasn’t possible to

measure higher voltages (it has a maximum range limited to the interval between -30 V and 30 V).

Therefore we were not able to determined the value of V1/2 .

This last result shows some of the greatest advantages of using arrays of MTJs:

- with relatively small bias currents high output voltages are achieved;

- the higher robustness of a MTJ array which tolerate a very high potential drop

between its terminals without disruption and with just a small TMR drop.

Regarding the robustness it can be better understand when compared with the result for an

individual MTJ:

Figure 7.15: TMR of a MTJ decreasing with the applied voltage.

The figure above[17]

shows a decreasing of 50% in TMR for an individual junction biased with

350 mV, which is substantially less voltage than what was applied to the MTJ array, which supported

30V without suffering a 50% TMR loss. This feature makes the MTJs series particularly useful and

usable since they support any electrostatic discharge that could occur in their handling, and is justified

by the “distribution” of the total potential drop in the series by each of its individual elements.

62

M

H

H

Functionalized stripe

Nanoparticle

While the work of characterization of the sensors was being done, one set of sensors was sent

to another partner institution in the project with the goal of proceeds real applications measurements.

Thus, sensors of the first generation (ZarMTJ1), were sent. The wheel device already explained was

used.

Figure 7.16: Graph - Response of one sensor under magneto nanoparticles excitation. Scheme - Functionalized stripe with nanoparticles attached, and a vertically magnetized nanoparticle.

During this experiment the particles where vertically magnetized, as shown in the previous

figure. The wheel was not rolling freely but fixed and was pulled over the sensor When the

nanoparticles are trapped in the stripe they have a certain tendency to accumulate themselves at the

edge of the stripes. Thus, the boundaries of the stripes have a higher concentration of the

nanoparticles, and will create a higher magnetic field in these regions of the stripes. Therefore, when

the stripe is pulled over the sensors, considering this nom-uniformity in the vertically magnetized

particles concentration over the stripe, two spikes appears close to the boundaries of the stripe (Figure

7.16). Therefore the shape of the curve is concordant with the prediction made in Chapter 3.

Stripe width

63

8. Conclusions and Future work

During this thesis several samples of MTJs connected in series were patterned. TMRs off about

70% were achieved as well as samples with sensibilities of 1.2%/Oe. It was proved the great

robustness of these devices from which we can get a Voltage output of the order of some tenth of

Volts without occur their disruption.

Being an ongoing work it still requires a fully noise characterization (Sv vs #junctions , Sv vs

Vbias) and the pattern of new samples with improved TMR singnal and Sensibility.

64

65

Bibliography

[1] A. B. Henriques, J. C. Romão, electromagnetismo, IST Press, 2006. [2] S.O. Kasap, Principles of Electronic Materials and Devices, McGraw-Hill, 2nd edition, 2006. [3] R.Waser, Nanoelectronics and information technology: advanced electronic materials and

novel devices,Wiley-VCH, 2nd edition, 2005. [4] A. Fert, et al., Giant Magnetoresistance of (001)Fe/(001)Cr Magnetic Superlattices, Phys. Rev.

Letter 61(21): 2472-2475, 1988. [5] B. Dieny, et al., Giant magnetoresistance in soft ferromagnetic multilayers, Phys. Rev. B 43(1):

1297-1300, 1991. [6] M. Julliere, Tunneling between ferromagnetic films, Phys. Lett. A 54: 225, 1975. [7] P. Grunberg, et al., Enhanced Magnetoresistance in layered magnetic structures with

antiferromagnetic interlayer exchange, Phys. Rev. B 39(7): 4828-4830, 1989. [8] J. C. Slonczewski, Conductance and exchange coupling of two ferromagnets separated by a

tunneling barrier, Phys. Rev. B 39: 6995, 1989. [9] D. Wang, et al., Spin dependent tunneling junctions with reduce Neel coupling, J.Appl.Phys.,

93(10),2002. [10] P. Bruno, C. Chappert, Giant Magnetoresistance of (001)Fe/(001)Cr Magnetic

Superlattices,Phys. Rev. Letter 67: 1602, 1991. [11] P.Wisniowski, et al., Effect of free layer thickness and shape anisotropy on the transfer curves

of MgO magnetic tunnel junctions, J.Appl.Phys., 103, 2008. [12] Manuel João de Moura Dias Mendes, Micromagnetic Simulations of Spin Valve devices,

Senior thesis,Instituto Superior Técnico, 2005. [13] Susana Isabel Pinheiro Cardoso de Freitas, Dual-Stripe GMR and Tunnel Junction Read

Heads and Ion Beam Deposition and Oxidation of Tunnel Junctions, PhD thesis, Instituto Superior Técnico, 2001.

[14] Ricardo Alexandre de Matos Antunes Ferreira, Ion Beam Deposited Magnetic Spin Tunnel

Junctions targeting HDD Read Heads, Non-volatile Memories and Magnetic Field Sensor Applications, PhD thesis, Instituto Superior Técnico, 2008.

[15] Jianguo Wang, Low-resistance tunnel junctions for read head applications, PhD thesis,

Instituto Superior Técnico, 2002. [16] Haohua Li, Spin valve read elements and sensors, PhD thesis, Instituto Superior Técnico,

2004. [17] Filipe Arroyo Cardoso, Design, optimization and integration of magnetoresistive biochips, PhD

thesis, Instituto Superior Técnico 2010. [18] Ricardo Alves Ferreira Costa e Sousa, Magnetic Random Acess Memory (MRAM) based on

Spin Dependent Tunnel Junctions, PhD thesis, Instituto Superior Técnico [19] R. Guerrero, et al., Low frequency noise in arrays of magnetic tunnel junctions connected in

series and parallel, Appl. Phys., Vol. 105, 2009.

66

67

A. Appendix

Figure .A.1: Mask use – Blue lines: top contacts; Yellow lines: define the shape of the contact pathways in the last passivation layer; in the lower right corner: two individual junctions, with a

reference role for comparison.

68

148µm

956µm

Figure A.2: Top: Series detail – Green: bottom contacts; Red: pillar junctions; Blue: top contacts Bottom: The entire array of MTJs, with intermediate contacts;.

Figure A.3: 3D scheme of MTJ’s connected in serious: the current coming from a top contact goes through a junction, in a descending direction until reach the bottom contact; flows in the bottom contact reaching the other junction and crosses it in the upper direction; now the current flows again in a top contact until the next junction and so on. The top contact, the bottom contact and the pillar junction correspond to the blue lines, green lines and red lines in the previous figure

69

B. Appendix

Run Sheet for MTJs Array fabrication: Influenza_ 2010

Process Start: / /2010 Process Complete: / /2010

Substrate: Si wafer / 800 Å SiO2 : ~1 inch2

Machine: Nordiko 2000

Magnetic Tunnel Junction Structure:

Bottom electrode: Ta 50 Å / Ru 180 Å / Ta 30 Å / PtMn 200 Å / CoFe 20 Å / Ru 9 Å / CoFeB 30 Å

Oxide Barrier: MgO x Å

Top electrode: CoFeB y Å / Ru 50 Å / Ta 50

Total height: 619 + x + y Å

Calibration samples: VSM

Top electrode for 2nd etch (glass substrate)

Machine sequences:

Seq. 3: Pre-sputtering of all targets x 2;

Seq. 38: Xianghiong

Seq. 39: FILIPE

Caracterization: VSM after annealed

STEP 1 Tunnel Junction Deposition

70

Step 1: Read Values

B.P:8.6x10-8 Torr

Seq. 3: pre sputtering of all targets x2

B.P:8.4x10-8 Torr

Sequence:

(#/name) Function : Read Values

38/Xianghiong

F18, Ta 50 40 mA | 331V | 10 W | 9.7 sccm | 4.6 mTorr | S4T3 | 100%

F7, Ru 180 40 mA | 302V | 10 W | 7.8 sccm | 5.1 mTorr | S4T2 | 100%

F6, Ta 30 40 mA | 329V | 10 W | 9.8 sccm | 4.6 mTorr | S4T3 | 100%

F5, PtMn 200 30 mA | 292V | - | 9.7 sccm | 4.6 mTorr | S4T3 | 100%

F9, CoFeB 20 F34 R3 B271 | 7.7 sccm |5.2 mTorr | S4T1 | 100%

F3, Ru 9 40 mA | 292V | 10 W | 7.8 sccm | 7.8 mTorr | S4T2 | 100%

F47, CoFeB 30 40 mA | 417V | 10 W | 8.7 sccm |5.2 mTorr | S4T4 | 100%

F94, MgO

Cleaning 2´30’’ F149 R0 B291 | 9.6 sccm |5.1 mTorr | S4T2 | 100%

39/FILIPE

F51, MgO x F129 R0 B272 | 9.5 sccm |8.2 mTorr | S4T5 | 50%

F50, CoFeB y 40 mA | 418V | 10 W | 8.8 sccm |5.1 mTorr | S4T4 | 100%

F4, Ru 50 40 mA | 302V | 10 W | 7.8 sccm | 5.1 mTorr | S4T2 | 100%

F18, Ta 50 40 mA | 330V | 10 W | 9.9 sccm | 4.6 mTorr | S4T3 | 100%

71

Machine: Nordiko 7000

Machine sequence:

Seq.17 – mod.2 – f.9 (1’ soft sputter etch) P=60W/40W, p=3mTorr, 50 sccm Ar

mod 3 – f.19 (150A TiW, 27’’) P=0.5 kW, 3mTorr, 50sccm Ar + 10 sccm N2

Readings – Module 2

Run# Power1 Power2 Ar flux (sccm) Pressure (mTorr)

F59 R8 B131 F40 R1 50.2 3.1

Readings – Module 3

Run# DC Power (kW) Voltage (V) Current (A) Ar flux (sccm) Pressure (mTorr)

0.50 419 1.2 50.5 3.0

Calibration samples: Top electrode for 2nd etch

Machine: Vapour prime; SVG Track; DWL.

Vapour Prime: Program 0 - 30 min.

Coating: Recipe 6/2 - 1.5 μm photoresist

Masks: a)ZarMTJ1 Map: HUGO ( Die size: 8.2x8.2 mm2)

b)Zaragoza_2x30_L1

Alignment mark: none

Energy: 80

Power: 120mW

Focus: -50

Develop : Recipe 6/2 Development time : 1 min

STEP 3 1st Lithography – Main Pillar Definition

STEP 2 MTJ Passivation – 150 Å TiWN2

Easy

Axis

a b

b b

X,Y

72

Optical Inspection:

On an optical microscope, with a green filter to confirm the sample development (filters the wave

lengths which impress the photoresist, allowing to develop the sample a little more if needed).

Sample Comments

All samples Every time that one sample was not perfectly exposed the optical lithography was repeated.

Machine: N3600

Total thickness to etch: 619 + x + y Å (etch rate: ~1 Å/s time: 619 + x + y s)

Base Pressure (Torr): 2.04x10-7 T Cryo (K): 107

Batch: etchjunction

Recipe etch junction stack all

Steps: etch pan 60 deg

cool_down_200s

Assist Gun: 160W 105mA +735V/-350V 12sccm Ar; Assist Neut: 30% subst.rot 60˚ subst.pan

Assist Gun Power (W) V+ (V) I+ (mA) V- (V) I- (mA) Ar Flux (sccm)

Read Values 195 723.5 104.5 344.3 2.6 10.6

STEP 4 1st Ion Milling – Total Structure Etch

73

Machine: Chemical Wetbench: Hot bath; ultrasounds machine.

Hot μ-strip + ultrasonic

Started:_____________ Stoped:_____________

Total Time in hot µ-Strip : 5h Ultrassonic Time : 40 min.

Rinse with IPA + DI water + dry compressed air

Optical Inspection:

Sample Comments

All Samples Some bottom contact remain shorted: problem will be solved in the second ion milling.

Machine: Vapour prime; SVG Track; DWL.

Vapour Prime: Program 0 - 30 min.

Coating: Recipe 6/2 - 1.5 μm photoresist

Masks: a)ZarMTJ2 Map: HUGO ( Die size: 8.2x8.2 mm2)

b)Zaragoza_2x30_L2

Alignment mark: (168, 54) µm

Energy : 80

Power : _120mW

Focus : -70

Develop : Recipe 6/2 Development time : 1 min

STEP 5 Resist Strip

STEP 6 2nd Lithography – Junction Definition

Easy

Axis

a b

b b

X,Y

74

Optical Inspection:

Machine: N3600

Total thickness to etch: 309+x+y Å (etch rate: ~1 Å/s time: 309+x+y s)

Base Pressure (Torr): 2.12x10-7

Batch: etchjunction

Recipe etch junction 60/30

Steps: Etch Junction Top electrode

cool_down_200s

etch pan 30 deg

Assist Gun: 160W 105mA +735V/-350V 12sccm Ar; Assist Neut ; 30% subst.rot 30º subst.pan

Calibration Sample Structure

CoFe 20Å | Ru 9Å | MgO xÅ | CoFeB yÅ | Ru 50Å | Ta 50Å | 150 TiWN2

Assist Gun Power (W) V+ (V) I+ (mA) V- (V) I- (mA) Ar Flux (sccm)

Read Values 195 723.3 104.6 350 2.5 10.2

Wafer samples Etching Turn Time Effect

Etch pan 30˚ 59’

Etch pan 60˚ 158’

Etch pan 30˚ 59’ Calibration sample tranlucid

STEP 7 2nd Ion Milling – Top Electrode and Junction Definition

75

Responsible: Fernando

Machine: UHV2

Dep. Rate ~11.43 Å/s

Deposition

Time

Al2O3 thickness Ar gas flow Pressure Power Source

62’ 700 A 45sccm 200W

Comments: Calibration sample: profilometer and ellipsometer.

Machine: Chemical Wetbench: Hot bath; ultrasounds machine.

Hot μ-strip + ultrasonic

Started:_____________ Stoped:_____________

Total Time in hot μ-strip : 3 days Ultrasonic Time : 40 min.

Rinse with IPA + DI water + dry compressed air

Optical inspection:

STEP 8 Insulating Layer Deposition

STEP 9 Oxide Lift-Off

76

Machine: Vapour prime; SVG Track; DWL.

Vapour Prime: Program 0 - 30 min.

Coating: Recipe 6/2 - 1.5 μm photoresist

Masks: a)ZarMTJ3 Map: HUGO ( Die size: 8.2x8.2 mm2)

b)Zaragoza_2x30_L3

Alignment mark: (168, 54) µm

(168,168) µm

Energy : 80

Power : _120mW

Focus : -50

Develop : Recipe 6/2 Development time : 1 min

Optical Inspection:

Machine: Nordiko 7000

Machine sequence:

Seq.48 (svpad) – mod.2 – f.9 (1’ soft sputter etch) P=60W/40W, p=3mTorr, 50 sccm Ar

mod.4 – f.1 (1500A Al, 40’’) P=2 kW, 3mTorr, 50 sccm Ar

mod 3 – f.19 (150A TiW, 27’’) P=0.5 kW, 3mTorr, 50sccm Ar + 10 sccm N2

STEP 11 Contact Leads Deposition 1500 Å Al98.5Si1.0Cu0.5 / 150 Å TiWN2

STEP 10 3rd Lithography – Top Contact

Easy

Axis

a b

b b

X,Y

77

Readings – Module 2

Run# Power1 Power2 Ar flux (sccm) Pressure (mTorr)

F60 R8 B122 F40 R2 50.2 3.2

Readings – Module 4

Run# DC Power (KW) Voltage (V) Current (A) Ar flux (sccm) Pressure (mTorr)

2 404 5.0 50.5 3.1

Readings – Module 3

Run# DC Power (KW) Voltage (V) Current (A) Gas flux (sccm) Pressure (mTorr)

0.50 417 1.2 50.2 Ar / 10.6 N2 2.9

Machine: Chemical Wetbench: Hot bath; ultrasounds machine.

Hot μ-strip + ultrasonic

Started:_____________ Stoped:_____________

Total Time in hot μ-strip : 1 day Ultrasonic Time : 1 hour

Rinse with IPA + DI water + dry compressed air

Optical inspection:

Comment

This picture shows a localized defected resulting from a over development in the previous lithography.

STEP 12 Metal lift-off

78

Machine: Vapour prime; SVG Track; DWL.

Vapour Prime: Program 0 - 30 min.

Coating: Recipe 6/2 - 1.5 μm photoresist

Masks: a)ZarMTJ4 Map: HUGO ( Die size: 8.2x8.2 mm2)

b)Zaragoza_2x30_L4

Alignment mark: (168, 54) µm

(168,168) µm

(168, 256) µm

Energy : 80

Power : _120mW

Focus : -50

Develop : Recipe 6/2 Development time : 1 min

Optical Inspection:

Responsible: Fernando

Machine: UHV2

Dep. Rate ~11.43 Å/s

Deposition

Time

Al2O3 thickness Ar flow (sccm) Pressure Power Source

1h27’ 1000 A 45 200 W

Comments: Calibration sample: profilometer and ellipsometer.

Machine: Chemical Wetbench: Hot bath; ultrasounds machine.

Hot μ-strip + ultrasonic

Started:_____________ Stoped:_____________

STEP 14 Final oxide deposition

STEP 15 Oxide lift-off ___

STEP 13 4th Lithography – junction contact opening

Easy

Axis

a b

b b

X,Y

79

Total Time in hot μ-strip : 1 day Ultrasonic Time : 30 min.

Rinse with IPA + DI water + dry compressed air

Optical inspection:

Machine: Old annealing setup

Temperature: 320˚C real - 280˚C equipment setpoint

Annealing time: 1h

Applied field: 4 kOe.

STEP 16 Magnetic Anneal