0 Bewertungen0% fanden dieses Dokument nützlich (0 Abstimmungen)

25 Ansichten58 SeitenF-Praktikum_INF12_RKM

© © All Rights Reserved

0 Bewertungen0% fanden dieses Dokument nützlich (0 Abstimmungen)

25 Ansichten58 SeitenF-Praktikum_INF12_RKM

Sie sind auf Seite 1von 58

Alexander Schwarz

Oktober 2011

Inhaltsverzeichnis

2. Prinzip der Rasterkraftmikroskopie

3. Vorbereitungsaufgabe

4. Graphit

5. Anfertigung des Protokolls

6. Zusätzliche Unterlagen zum Versuch

6.1. Originalpublikation: Atomic Force Microscope

6.2. Beiblatt: Prinzip eines LID Regelkreises

6.3. Übersichtsartikel: Nanoscale Imaging and Force Analysis

with Atomic Force Microscopy

6.4. Weiterführende Literatur

1. Aufgabenstellung und Ablauf

Der Ablauf untergliedert sich in (i) die häusliche Vorbereitung mit Hilfe (aber nicht ausschließlich)

dieser Mappe, (ii) die Durchführung des Versuchs, (iii) die Auswertung der Daten, (iv) das Anfertigen

des Protokolls und (v) die Nachbesprechung des Protokolls. Eine Gruppe pro Semester muss

außerdem einen Vortrag über den Versuch vorbereiten und halten. Die Qualität aller Teile ist

Bestandteil der Benotung.

Ziel dieses Versuchs ist es die Rasterkraftmikroskopie als Methode zur Untersuchung von

Oberflächen mit hoher Ortsauflösung kennen zu lernen und die Anknüpfungspunkte zur

Kontinuumsmechanik (Vorl. Physik I) und Festkörperphysik (Vorl. Physik IV) zu verstehen. Zur

Vorbereitung ist das Studium dieser Praktikumsmappe und der beiliegenden Literatur sowie die

Bearbeitung der Vorbereitungsaufgabe (s. Kap. 3) Voraussetzung. Abhängig vom Wissensstand kann

die Verwendung weiterer Quellen (Lehrbuch, Internet, …) notwendig sein. Zum Beispiel taucht in den

Unterlagen der piezoelektrische Effekt auf, ist aber nirgends im Detail erläutert. Wer nicht weiß,

worum es sich handelt, informiere sich.

Der tatsächliche Ablauf des Versuchs und die Schwerpunkte liegen bei jedem Versuch etwas anders.

Deshalb sollten keinesfalls sollten Teile des Protokolls vorgeschrieben werden. Das gilt auch für die

Theorie, selbst wenn diese Vorgehensweise für andere Versuche empfohlen wird. Stattdessen ist es

sehr hilfreich, sich während des Versuchs Notizen zu den besprochenen Aspekten zu machen, denn

diese Aspekte, die von Gruppe zu Gruppe variieren, sollen in dem anzufertigenden Protokoll betont

werden.

Der praktische Teil des Versuchs beginnt mit der Einführung in die Grundlagen der

Rasterkraftmikroskopie (s. Kap. 2 und Literatur in Kap. 6). Sie findet in Form eines Gesprächs statt, in

dem die physikalischen Prinzipien und die technischen Aspekte auf denen die Rasterkraftmikroskopie

aufbaut, mit eigenen Worten dargestellt werden sollen. Danach wird die Bedienung des

verwendeten Rasterkraftmikroskops besprochen. Die sich daran anschließenden Messungen dienen

dem Kennenlernen der Funktionsweise eines Rasterkraftmikroskops (Betriebsmodi, Regelkreis,

Spektroskopie usw.). Im weiteren Verlauf werden dann selbstständig verschiedene Probensysteme

untersucht, wobei Aspekte wie Auflösungsvermögen oder der Zusammenhang zwischen

Oberflächenstruktur und anderen Eigenschaften der Probe (z. B. optischen Eigenschaften oder

Anwendung bzw. Funktion der Oberfläche) untersucht werden sollen. Welche Proben genau

untersucht werden, variiert von Gruppe zu Gruppe. Die ersten Messungen werden allerdings immer

auf Graphit durchgeführt (s. Kap. 4). Nach dem Pflichtprogramm besteht Gelegenheit nach

Absprache eigene Proben zu untersuchen. Der Phantasie sind da keine Grenzen gesetzt, allerdings

können zu große Strukturen mit dem vorhandenen Mikroskop nicht untersucht werden. Bei Interesse

endet der praktische Teil des Versuchs mit einer kleinen Laborführung. Das Protokoll ist bis zum an

die Praktikumswoche anschließenden Freitag abzugeben. Näheres zum Protokoll ist in Kap. 5

beschrieben.

2. Prinzip der Rasterkraftmikroskopie

Die Rasterkraftmikroskopie (engl. atomic force microscopy; AFM) gehört zu den

Rastersondenmikroskopie-Techniken (engl. scanning probe microscopy; SPM) und beruht auf

einem einfachen mechanischen Verfahren zum punktweisen Abtasten einer

Probenoberfläche. Dazu wird der Kraftsensor (engl. force probe; s. Abb. 1), der aus einem

einseitig eingespannten Federbalken (engl. cantilever) mit einer scharfen Spitze am freien

Ende (engl. tip) besteht, zeilenweise über eine Probenoberfläche gerastert (engl. scanning,

deshalb auch scannig force microscopy; SFM). Die XYZ-Rastereinheit besteht meist aus

einem piezoelektrischen Material oder basiert, wie in dem hier verwendeten Mikroskop, auf

dem Lautsprecherprinzip. Während der Rasterbewegung wird die Auslenkung des

Federbalken detektiert, wobei verschiedenste optische und elektrische Verfahren

gebräuchlich sind (s. Literatur in Kap. 6). Das hier benutze Mikroskop verwendet das

Lichtzeigerprinzip. Abbildung. 2 zeigt den schematischen Aufbau eines Rasterkraft-

mikroskops mit den wesentlichen Komponenten.

dem Substrat, an dem der Federbalken mit integrierter Spitze angebracht ist. Die Abmessungen sind

im Bild angegeben. Das am Rand sichtbare Substrat hat Dimensionen im mm-Bereich, damit der

Kraftsensor mit einer Pinzette in das Mikroskop eingebaut werden kann.

Messmodi

Der Federbalken kann sowohl im statischen also auch im dynamischen Modus eingesetzt

werden. Im statischen Modus (engl. static mode oder dc-mode) wird direkt die statische

Verbiegung aufgrund der unterschiedlichsten Kräfte zwischen Spitze und Probe detektiert.

Im dynamischen Modus (engl. dynamic mode oder ac mode) oszilliert der Federbalken

selbstangeregt oder erzwungen. Aus dem oszillierenden Signal am Detektor kann entweder

die Frequenzänderung (engl. frequency modulation; FM-AFM), die Amplitudenänderung

(engl. amplitude modulation; AM-AFM) oder die Phasenänderung (engl. phase modulation;

PM-AFM) aufgrund der Spitze-Probe-Wechselwirkungen bestimmt werden. Des Weiteren

kann die Spitze im Kontakt mit der Probe sein (engl. contact mode) oder nicht (engl. non-

contact mode; NC-AFM). Im dynamischen Modus kann die Spitze auch nur im Bereich des

unteren Umkehrpunktes mit der Probe im Kontakt sein (engl. tapping mode; TM-AFM).

In der Literatur ist die verwendete Nomenklatur für die verschiedenen Messmodi

inkonsistent, weil sie historisch gewachsen ist. So wird der static mode oft mit dem contact

mode gleichgesetzt, obwohl bei langreichweitigen Kräften die statische Verbiegung auch

ohne Probleme im Nichtkontaktbereich detektiert werden kann (und wird). Andererseits

wird FM-AFM und NC-AFM oft synonym benutzt, obwohl Ersteres eine Information über den

Detektionsmechanismus enthält, nämlich dass die Frequenzverschiebung gemessen wird,

während Letzteres etwas über den Spitze-Probe-Abstand im Messbetrieb aussagt. Mit dem

hier verwendeten Mikroskop wird Und tatsächlich kann man die Freqzuenzverschiebung

auch detektieren, wenn die Spitze im unteren Umkehrbereich die Oberfläche berührt. Im

Rahmen dieses Versuchs wird nur im Kontaktbereich mit dem statischen Modus und dem

Amplitudenmodulationsmodus gearbeitet.

Rastereinheit und Regelkreis. Die Datenaufnahme wird über ein spezielles Computerprogramm

gesteuert, das auch die Darstellung und Auswertung der Daten erlaubt.

Regelkreis

Alle Messmodi können sowohl im ungeregelten als auch im geregelten Modus betrieben

werden. Ohne Verwendung des Regelkreises wird direkt das detektierte Signal S vom

Federbalken (je nach Modus also die Verbiegung, die Frequenzverschiebung, die

Amplitudenänderung oder die Phasenänderung) in ortsabhängig aufgezeichnet, also S(X,Y).

Mit Regelkreis (s. Literatur in Kap. 6.2) wird das Z-Stellelement der XYZ-Rastereinheit so

nachgeregelt, das S konstant bleibt. Die ortsabhängige Z-Position des Z-Stellelements, also

Z(X,Y) wird üblicherweise als Topographie bezeichnet, da diese Abbildung der

Oberflächenstruktur entspricht.

Spektroskopie

Im Spektroskopiemodus wird die XY-Rasterung und der Regelkreis ausgeschaltet und das

detektierte Signal S in Abhängigkeit von der Z-Position des Z-Stellelements, also S(Z), an

einer bestimmten XY-Position aufgezeichnet. Im statischen Modus wird diese Spektroskopie

Kraft-Distanz-Kurve genannt, da bei bekannter Federkonstante k aus der Auslenkung Δz des

Federbalkens (also dem im statischen Modus detektierten Signal S) die Kraft F berechnet

werden kann (s. Kap. 3).

Das Mikroskop selbst wird über einen Rechner mit einem speziellen Programm gesteuert,

das auch die Daten aufzeichnet und eine Auswertung der Daten erlaubt. Dabei muss

beachtet werden, dass mikroskopseitig die Daten als analoge Signal vorliegen und auch die

Ansteuerung nur über analoge Signale möglich ist, während die Software des Rechners nur

digitale Signale verarbeiten und ausgeben kann.

3. Vorbereitungsaufgabe

Aufgenommen wird die Verbiegung des Federbalkens Δz in Abhängigkeit von der Position

des Z-Stellelementes. Mit Hilfe der Federkonstanten k kann aus der Verbiegung die wirkende

Kraft bestimmt werden. Weit entfernt von der Probe ist die Spitz-Probe-Wechselwirkung

verschwindend gering. Der Federbalken ist nicht verbogen (Nulllinie). Fährt das Z-

Stellelement Richtung Oberfläche beginnt der Federbalken sich zu verbiegen und an einem

bestimmten Punkt springt die Spitze auf die Oberfläche (Einspringpunkt). Die Spitze berührt

jetzt die Oberfläche und der Federbalken wird entsprechend der Vorwärtsbewegung des Z-

Stellelementes nach oben gebogen. Wird das Z-Stellelement in die andere Richtung bewegt,

verringert sich die Verbiegung des Federbalkens bis er wieder gerade ist (Nulllinie). Die

Spitze ist aber immer noch im Kontakt mit der Oberfläche und wird jetzt nach unten

gebogen, wenn das Z-Stellelement sich noch weiter von der Oberfläche weg bewegt. Erste

an einem bestimmten Punkt löst sich die Spitze sprunghaft von der Oberfläche (Abrisspunkt)

und ist wieder gerade (Nulllinie). Da Hin- und Rückweg des Z-Stellelementes nicht identisch

sind, spricht man von einer Hysterese.

Federbalkens (Ordinate) kann direkt die Kraft berechnet werden.

Einsprung- und Abrisspunkt befinden sich an unterschiedlichen

Positionen des Z-Stellelementes (Hysterese zwischen Hin- und

Rückweg).

Der Kurvenverlauf und das Auftreten einer Hysterese soll verstanden werden, in dem man

sich die Potentiallandschaft anschaut, in der sich die Spitze befindet. Sie ist durch die

potentielle Energie des Federbalkens und die Spitze-Probe-Wechselwirkung gegeben. Das

Federpotential ist bekannt. Als Modelpotential für die Spitze-Probe-Wechselwirkung kann

das Lennard-Jones Potential verwendet werden:

𝑟0 12 𝑟0 6

𝐸LJ (𝑧) = 𝐸0 �� � − 2 � � �

𝑧 𝑧

Für verschieden Positionen u (z. B. u = 1 nm, 0.5 nm, 0.2 nm, 0.1 nm, 0.05 nm) des Z-

Stellelementes soll das Federpotential, das Lennard-Jones-Potential und die Gesamtenergie

E(z,u) grafisch dargestellt werden. Dabei ist u der Abstand der Oberfläche zur Ruhelage des

Federbalkens bzw. der Spitze und z die tatsächliche Position des Federbalkens bzw. der

Spitze. Als Parameter für das Federpotential soll k = 1 N/m verwendet werden und für das

Lennard-Jones-Potential E0 = 1 eV und r0 = 0.3 nm.

Aufgabe: Wie kann aus dieser Energiebetrachtung die Hysterese in den Kraft-Distanz-

Kurven erklärt werden?

4. Graphit

Graphit ist eine reine Kohlenstoffmodifikation deren atomare Struktur in Abb. 4 skizziert ist.

Die wabenförmig angeordneten Kohlenstoffatome in der Ebene sind sp2-hybridisiert und

bilden zu je drei nächsten Nachbarn kovalente Bindungen mit dem Abstand a = 142 pm. Die

einzelnen Ebenen sind über van der Waals Wechselwirkungen mit der Stapelfolge ABA

gebunden. Das nicht an der kovalenten Bindung beteiligten p-Elektronen bilden das

Leitungsband und sorgen für gute Wärmeleitfähigkeit sowie den metallischen Glanz. Auf

Grund der schwachen Bindung zwischen den Ebenen lassen sich saubere Oberflächen durch

Spalten mit einem Tesafilmstreifen herstellen. Nach dem Spaltvorgang bilden sich auf der

Oberfläche zahlreiche atomar flache Terrassen, die durch Stufen voneinander getrennt sind.

Die Stufenhohen entsprechen ganzzahligen Vielfachen des Ebenabstandes c/2 = 335 pm.

Abb. 4: Struktur von Graphit. Links ist die Aufsicht zu sehen, die die Symmetrie der Oberfläche zeigt. Die

gestrichelten Linien deuten die Lage der Atome in der darunter liegenden Ebene an. Rechts ist die

lagenweise Anordnung der ABA-Stapelung zu sehen.

In dem Protokoll sollen die mit dem Rasterkraftmikroskop durchgeführten Experimente in

einer wissenschaftlichen Darstellungsweise schriftlich niedergelegt werden. Die

unterschiedlichen Aspekte und untersuchten Probensystem verbieten einen

Standardaufbau wie z. B. im Anfängerpraktikum (Einleitung, Theorie, Durchführung,

Auswertung, Diskussion, … oder so ähnlich)! Es sollte eine in sich geschlossenen Darstellung

des gesamten Versuchs sein, wobei inhaltliche Zusammenhänge wichtiger sind als der

zeitliche Ablauf während des Praktikums. Wichtig ist ein roter Faden in der Darstellung und

den Argumentationsketten. Abbildungen und Skizzen - sie sagen oft mehr als viele Worte -

sollen mit dem Text Hand in Hand gehen. Die Bildunterschriften sollten knapp sein und keine

langatmige Wiederholung des Haupttextes. Nicht alle aufgenommenen Daten müssen als

Abbildung im Protokoll auftauchen, sondern die grafische Darstellung der Messdaten soll die

Argumentation beispielhaft unterstützen. Im Text muss auf die Messdaten und die

dazugehörigen Abbildungen Bezug genommen werden. Der Hauptteil des Protokolls sollte

sich mit euren Daten beschäftigen und nicht mit der Einleitung oder Theorie der

Rasterkraftmikroskopie.

Die Ausfertigung des Protokolls sollte sich nicht an anderen Protokollen richten, sondern den

individuellen Versuchsablauf widerspiegeln. Er ist von Gruppe zu Gruppe unterschiedlich,

nicht nur was die untersuchten Probensysteme angeht, sondern auch die

Schwerpunktsetzung der unterschiedlichsten Aspekte der Rasterkraftmikroskopie. Dazu ist

es sinnvoll auf eure Aufzeichnungen zurückzugreifen.

Das Kopieren von Textpassagen aus anderen Protokollen wird als Betrugsversuch gewertet

(ja - ich führe Plagiatstests durch), ebenso wie die Verwendung von fremden Messdaten.

Der Versuch gilt dann als nicht bestanden. Bitte nur selbst angefertigte Abbildungen

verwenden. Das bezieht sich nicht nur auf Messdaten, sondern auch auf Grafiken, Skizzen

und Schemata aus Büchern, der Vorbereitungsmappe, aus dem Internet oder aus anderen

Protokollen. Auch mit Quellenangabe werde ich „Copy & Paste“ Abbildungen nicht

akzeptieren. Grund: Abbildungen sind ein wichtiger Bestandteil der Darstellung

physikalischer Sachverhalte und die Anfertigung sinnvoller Skizzen soll dadurch geübt

werden. Sie selbst anzufertigen führt außerdem zu einem nachhaltigeren

Verinnerlichungsprozess des Inhalts einer Abbildung als die „Copy & Paste“ Methode.

Schickt mir das Protokoll per Email, falls ihr es nicht handschriftlich anfertigt. Bitte

verwendet Dateinamen mit folgender Struktur: Nachname1Nachname2RKM.pdf

Das Protokoll darf nicht mehr als 20 Seiten umfassen (inkl. Anhängen etc.). Längere

Protokolle werden ungesehen zurückgegeben. Legt also den Schwerpunkt auf die

Experimente, die ihr gemacht habt, und nicht auf die Einleitung oder die Theorie. Der Stil

sollte sachlich und im Ausdruck präzise sein. Keinesfalls gehören persönliche Kommentare

oder ein „Fazit“ in das Protokoll. Ich lege Wert auf gute Lesbarkeit und berücksichtige sie bei

der Bewertung. Schreibt also verständlich, vermeidet Anglizismen und achtet auf eure

Rechtschreibung. Ich bewerte ausschließlich das erste abgegebene Protokoll.

oder “z (m)“ aber nicht „z [m]“, auch wenn das oft zu sehen ist. Bitte SI Einheiten

verwenden.

2. Haupttext in Schriftgröße 11 oder 12.

3. Bildbeschriftungen in etwas kleinerer (Schriftgröße 10). Bildbeschriftungen enden mit

einem Punkt und sollten knapp gehalten werden. Im Haupttext muss auf die

Abbildung verwiesen werden.

4. Formelzeichen kursiv, Einheiten gerade und Leerzeichen einbauen also „E = 1 eV“.

Tiefgestellte Formelzeichen oder laufende Indizes ebenfalls kursiv, aber nicht Kürzel

wie das LJ in ELJ(z) für Lennard Jones.

VoLUME 56, NUMBER 9 PHYSICAL REVIEW LETTERS 3 MAR. cH 1986

G. Binnig"~ and C. F. Quate' '

Edward L Gi.nzton Laboratory, Stanford University, Stanford, California 94305

and

Ch. Gerber"

IBM San Jose Research Laboratory, San Jose, California 95193

I, Received 5 December 1985)

The scanning tunneling microscope is proposed as a method to measure forces as small as 10

N. As one application for this concept, we introduce a new type of microscope capable of investi-

gating surfaces of insulators on an atomic scale. The atomic force microscope is a combination of

the principles of the scanning tunneling microscope and the stylus profilometer. It incorporates a

probe that 0does not damage the surface. Our preliminary

0

results in air demonstrate a lateral resolu-

tion of 30 A and a vertical resolution less than 1 A.

%e are concerned in this paper with the measure- lometer (SP)8 9 has been developed into a powerful

ment of ultrasmall forces on particles as small as single microscopic technique. Teague et al. ' have used the

atoms. %e propose to do this by monitoring the elas- SP to record three-dimensional images of surfaces with

tic deformation of various types of springs with the a lateral resolution of 1000 A and a vertical resolution

scanning tunneling microscope (STM). ' It has been a of 10 A. A related technique is the scanning capaci-

common practice to use the displacement of springs as tance microscope described by Matey and Blanc. "

a measure of force, and previous methods have relied They report a lateral resolution of 5000 A and a verti-

on electrostatic fields, magnetostatic fields, optical cal resolution of 2 A.

waves, and x rays. Jones~ has reviewed the devices The SP has much in common with the STM. The

that use variable capacitances and he reports that dis- tip in the STM and the stylus in the SP are both used

placements of 10 4 A can be measured. SQUIDs3 are to scan the surface, sense the variations of the sample,

superconducting elements that measure the expulsion and generate three-dimensional images. The stylus in

of magnetic fields in variable-inductance devices. the profilometer is carried by a cantilever beam and it

They are used in gravity gradiometers to measure dis-

placements of 10 6 A. Tabor and co-workers in their surface can be plastically deformed. "

rides on the sample surface. This means that a rough

The radius of

work with van der Waals forces have used optical in- this stylus is about 1 p, m, and the loading force ex-

terference methods to measure displacements of 1 A. tends from 10 to 10 N. '

With an x-ray interferometer constructed from a single The spring in the AFM is a critical component. %e

crystal of silicon, Deslattes' has also measured dis- need the maximum deflection for a given force. This

placements of 10 A which is about 1'/0 of the nuclear requires a spring that is as soft as possible. At the

diameter. same time a stiff spring with high resonant frequency

We are proposing a new system wherein the STM is is necessary in order to minimize the sensitivity to vi-

used to measure the motion of a cantilever beam with brational noise from the building near 100 Hz. The

an ultrasmall mass. The force required to move this resonant frequency, fo, of the spring system is given

beam through measurable distances (10 A) can be by f0= (I/2sr)(k/nto)', where k is the spring con-

as small as 10 ' N. The masses involved in the other stant and ttto is the effective mass that loads the spring.

techniques are too large to reach this value. This level This relation suggests a simple way out of our dilem-

of sensitivity clearly penetrates the regime of inter- ma. As we decrease k to soften the spring we must

atomic forces between single atoms and opens the door also decrease mo to keep the ratio k/mo large. The

to a variety of applications. The atomic force micro- limiting case, illustrated in Fig. 1, is but a single atom

scope (AFM) is a new tool designed to exploit this lev- adsorbed at site A in the gap of an STM. It has its own

el of sensitivity. It will be used to investigate both mass and an effective k that comes from the coupling

conductors and insulators on an atomic scale. %e en- to neighboring atoms.

vision a general-purpose device that will measure any The mass of the spring in manmade structures can

type of force; not only the interatomic forces, but elec- be quite small but eventually microfabrication'4 will be

tromagnetic forces as well. employed to fabricate a spring with a mass less than

%ith the STM, 6 the atomic surface structure of con- 10 '0 kg and a resonant frequency greater than 2 kHz.

ductors is well resolved. For bulk insulators7 an Displacements of 10 A can be measured with the

equivalent method is missing although the stylus profi- STM when the tunneling gap is modulated. The force

VoLUME 56, NUMBER 9 PHYSICAL REVIEW LETTERS 3 M~RcH 1986

E3

SCANNERS, A C

FE E DBACK

F FEEDBACK D ST& F

T ATOM AFM

BLOCK (ALUMINUM) Z

Q

Y

FIG. 1. Description of the principle operation of an STM B: AFM DIAMOND TIP DIAMOND

as well as that of an AFM. The tip follows contour B, in one C: STM TII (Au) TIP

case to keep the tunneling current constant (STM) and in LI: CANTILEVER, .8 mm

STM SAMPLE

the other to maintain constant force between tip and sample LEVER

E: MODULATING PIEZO

(AFM, sample, and tip either insulating or conducting). {Au- FOIL)~

F: VITON

The STM itself may probe forces when a periodic force on

the adatom 3 varies its position in the gap and modulates FIG. 2. Experimental setup, The lever is not to scale in

the tunneling current in the STM. The force can come from (a). its dimensions are given in (h). The STM and AFM

an ac voltage on the tip, or from an externally applied mag- piezoelectric drives are facing each other, sandwiching the

netic field for adatoms with a magnetic moment. diamond tip that is glued to the lever.

required to produce these displacements is 2&& 10 ' N In the first mode we modulated the sample in the z

and this is reduced by 2 orders of magnitude when a direction at its resonant frequency (5.8 kHz). The

cantilever with a 0 of 100 is driven at its resonant fre- force between the sample and the diamond stylus the —

quency. small force that we want to measure —

deflects the lev-

AFM images are obtained by measurement of the er holding the stylus. In turn, this modulates the tun-

force on a sharp tip (insulating or not) created by the neling current which is used to control the AFM-

proximity to the surface of the sample. This force is feedback circuit and maintain the force fo at a constant

kept small and at a constant level with a feedback level.

mechanism. %hen the tip is moved sideways it will In the second and third modes, the lever carrying the

follow the surface contours such as the trace 8 in Fig. diamond stylus is driven at its resonant frequency in

1. the z direction with an amplitude of 0. 1 to 10 A. The

The experimental setup is shown in Fig. 2. The can- force, fo, between sample and stylus changes the

tilever with the attached stylus is sandwiched between resonant frequency of the lever. This changes both

the AFM sample and the tunneling tip. It is fixed to a the amplitude and phase of the ac modulation of the

small piezoelectric element called the modulating tunneling current. Either of these can be used as a sig-

piezo which is used to drive the cantilever beam at its nal to drive the feedback circuits.

resonant frequency. In the fourth mode we used one feedback circuit. It

The STM tip is also mounted on a piezoelectric ele- was connected to the AFM and it was controlled by the

ment and this serves to maintain the tunneling current tunneling current in the STM. This system maintained

at a constant level. The AFM sample is connected to a the tunneling gap at a constant level by changing the

three-dimensional piezoelectric drive, i.e. , the x,y, z force on the stylus.

scanner. A feedback loop is used to keep the force The fourth mode was further improved by reconnec-

acting on the stylus at a constant level. Viton spacers tion of both feedback circuits in such a way that the

are used to damp the mechanical vibrations at high fre- AFM sample and the STM tip were driven in opposite

quencies and to decouple the lever, the STM tip, and directions with a factor n less in amplitude for the

the AFM sample. The tip is brought in close proximi- STM tip. The value of a ranged from 10 to 1000.

ty to the sample by mechanical squeezing of the Viton In contrast to previous methods, the absolute value

layers. High-frequency ( ) 100 Hz) filtering of build-

ing vibrations is done as in the pocket-size STM' with

of fo, the force on the stylus, was not well defined ex-

cept at the beginning of the measurement. The defor-

a stack of metal plates separated by Viton. mation of the spring, 4z, is we11 calibrated at the start-

We have operated the AFM in four different modes ing point, but as the measurement proceeds each com-

which relate to the connections of the two feedback ponent of the system moves in an unknown way be-

circuits, one on the STM and the other on the tip. All cause of thermal drifts. These change the initial cali-

four of these modes worked in principle. They each bration. Additionally, we know that the three-

served to maintain a constant force, fo, between the dimensional motion of the AFM sample must produce

sample and the diamond stylus while the stylus fol- modest amounts of change in Az so as to compensate

lowed the contours of the surface. for the simultaneous motion of the stylus as it follows

931

MBER 9

VOLUME 56, NUM PHYSICAL R EVIEW LETTER MARCH 1986

a senc

certaIn rang e that Is ep

e,

of the surface. . Therefore, even in

ln the

m

ifts the force fa will vary

b thth'"u h

s m eared out.

y

%ee su ev is determineded

suspect that this level

the force that Is nnecessary to pennetrate the film. n

the f'&rs t three methodds we use d s ma11 forces we 11

0f the surface and the a ue below the threshold an d there we lose the fine struc-

The fourth mode p roved to e ture.

b 1e. %'e used it to re cord the resu itss shown in FIgs. rovements overr the handmade

an the topograp h y for two differen t ion crease tthee resolution to t hee

sed h ere should increase

areas o Al 03) surface. Th we will be a e atomic

e x axis are displaced from m each other by microfabrication techniques s'4 will

d therma I dif r' 1o h ow us to reduce

a liow

' ' '

he vertical dash ed lines of Fig.. 3 m in d t th th unl 't by several or ers

e of magnitu d e.. %hen the instru-

variation in thee y direction o topog s that in an ultra Ig -v er

can e these w ere be well characterized we

'

the noise amplitu dec on the traces we now experience t a ll

can estimate lvc a eas 2 orders o f m agnitude. With

at least

' '

perio d i c corrugation these optimum conditions t h e thermally induced vi ra-

below j, A when t e tions er at room t e mperature will ill limit

d 100 A. t e to 10 N. If h s

cooled below 300 m

~ ~

films coverlllg the surfaces o t h e AFM. Th fi

water

to t e in teratomic forces. "I

es interesting w

In tth

when we compare it

'

ls

s tructure on the rihto

g te w1 I the binding energy is

onl be observed whe hen the app I'Ie d force exceeded the rials that are h e ld together with the

threshold. For smaller

sm forces tthee structure was wea If we arbitrarily equate

t h ee nergy to a force acting throug h a dt of06

A we findin that a bindingin energy o f 1 e V is equivalen

f rces therefore

m 0 ' N for ionic bonds to 0 li

—

M

some o f f tM

The limiting sensitivivit y of our instrum

30

ZO

IQ

traces for another

*

le. The curves grouped

ed under A were recorde d with

-p

. 3. The AFM trace s on a ceramic (Alq03) sample.

The vertical scale tran a the lowest to the ig set

tip off 10-" /A. For the l is near g

3x1010 8 N. The stabili t yo f the regulated orce i

10 N. Th e successive trac

traces are disp ace drift duced below this level. The force fo was rese

along the y axis. x10 Nfort e ra

value near 5x

932

VOLUME 56, NUMBER 9 PHYSICAL REVIEW LETTERS 3 M&RCH 1986

than these values. Therefore, we should be able to 4D. Tabor and R. H. S. Winterton, Proc. Roy. Soc. Lon-

measure all of the important forces that exist between don, Ser. A 312, 435 (1979); J. N. Israelachvili and D. Ta-

the sample and adatorns on the stylus. bor, in Progress in Surface and Membrane Sciencee, dited by

Finally, we want to point out that these forces also J. F. Danielli, M. D. Rosenberg, and D. A. Cadenhead

exist in the tunneling microscope itself and that they (Academic, New York, 1973), Vol. 7.

sR. D. Deslattes, Appl. Phys. Lett. 15, 386 (1968).

the STM. "'

can have a strong influence on the data collected with

The STM could be used as a force mi-

66. Binnig, H. Rohrer, Ch. Gerber, and E. Weibel, Phys.

Rev. Lett. 49, 57 (1982), and 50, 120 (1983).

croscope in the mode described here by simply mount- 7Thin insulating films can be studied with the STM as il-

ing the STM tip on a cantilever beam. lustrated in G. Binnig, H. Fuchs, J. Kubler, F. Salvan, and

We are pleased to thank J. Pethica for his inspiring A. R. Williams, to be published.

talk on the problems of the tip at the STM workshop SJ. B. P. W'illiamson, Proc. Inst. Mech. Eng. 182, 21

in Oberlech, Austria, which —although some months (1967).

—

later probably triggered the idea for the AFM. We 9K. H. Guenther, P. G. %ierer, and J. M. Bennett. Appl.

also want to express our appreciation to H. Rohrer and Optics 23, 3820 (1984).

~OE. C. Teague, F. E. Scire, S. M. Backer, and S. %'. Jen-

D. Pohl for those stimulating discussions. This work

was supported in part by a grant from the IBM Cor- sen, Wear 83, 1 (1982); see also P. A. Engel and D. B.

Millis, Wear 75, 423 (1982).

poration and in part by the Defense Advanced "J.R. Matey and J. Blanc, J. Appl. Phys. 57, 1437 (1985).

Research Projects Agency. 'zE. J. Davis and K. J. Stout, Wear 83, 49 (1982).

' T. Vorburger, private communication.

'4K. E. Peterson, Proc. IEEE 70, 420 (1982).

~sB. H. Flowers and E. Mendoza, Properties of Mauer (Wi-

'~Also at IBM San Jose Research Laboratory, San Jose, ley, London, 1970), Chap. 3, pp. 22-55.

Cal. 95193, and on leave from IBM Research Laboratory, ' H. Krupp, W. Schnabel, and G. Walter, J. Colloid Inter-

Zurich, Switzerland. face Sci. 39, 421 (1972).

~~~Also at Xerox Palo Alto Research Center, Palo Alto, '7J. H. Coombs and J. B. Pethica, IBM J. Res. Dev. (to be

Cal. 94304. published). These authors point out that the forces between

&'~on leave from IBM Research Laboratory, Zurich, the sample and the tip in the STM, especially when the tip

Switzerland. presses against an oxide layer, can be strong enough to alter

'G. Binnig and H. Rohrer, Sci. Am. 253, 50 (1985). the calibration of the z piezo,

2R. V. Jones, Proc. IEEE 17, 1185 (1970).

, E, C. Teague, Room Temperature Gold- Vacuum-Gold Tun-

3E. R. Mapoles, Development of Sauperconducting Gravity neling Experiments (University Microfilms International,

Gradiomerer for a Test of the Inverse Square Lavv (University Ann Arbor, Michigan, 1978), Chap. 4, pp. 141-148, where

Microfilm International, Ann Arbor, Michigan, 1981), p. 4; he discussed the van der Waals forces of attraction between

J. Clark, Physics (Amsterdam) 126BAC, 441 (1984). two gold spheres used in vacuum tunneling.

933

Prinzip eines LID-Regelkreises

Sollwert (SW)

Regler Regelstrecke

Istwert (IW)

𝑡 𝑑(𝑆𝑆−𝐼𝐼)

𝑆𝑆 = 𝑎 𝑆𝑆 − 𝐼𝐼 + 𝑏 ∫𝑡 𝑆𝑆 − 𝐼𝐼 d𝑡 + c

0 𝑑𝑡

LID bedeutet in diesem Zusammenhang linear, integral und differential, wie man sich an

Hand der Gleichung klar machen kann, in der a, b und c die entsprechenden Verstärkungs-

faktoren sind.

Vorüberlegung:

Was sind Sollwert, Istwert und Stellgröße bei einem Rasterkraftmikroskop?

1

1

Nanoscale Imaging and Force Analysis with Atomic Force

Microscopy

Hendrik Hölscher, André Schirmeisen, and Harald Fuchs

1.1

Abstract

The atomic force microscope (AFM) – somtimes also denoted as scanning force

microscope (SFM) – has developed into the most widespread tool for nanotech-

nology. This technique was invented by Binnig et al. [11] in 1986 and is nowa-

days used not only in physical, chemical, biological, and medical research lab-

oratories, but also in many companies for tasks such as product development

and routine quality control. Its success is due to the high resolution imag-

ing capabilities in combination with its versatility, making it possible to image

surfaces down to the atomic scale. Due to its universality, AFM can be ap-

plied to a large variety of samples and it can be adapted to many different

environments (gaseous, liquid, or vacuum).

The atomic force microscope can be operated in different modes, which can

be categorized into static and dynamic-modes. We give an introduction to the

basic concepts of these two modes and present applications under different

environmental conditions. Some practical issues for the experimentalist are

pointed out and special care is taken to explain the quantitative relation be-

tween experimental parameters and physical magnitudes.

1.2

Principles of Atomic Force Microscopy

1.2.1

Basic Concept

has been a challenge for scientists for many years. In fact only very recently

a demonstration was given that these forces can be determined for a single

atomic bond, by the powerful technique of atomic force microscopy. But how

Copyright © 2006 Wiley-VCH, Weinheim (Germany)

ISBN: 3-527-XXXXX-X

2 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

billionth of one Newton?

The answer to this question is surprisingly simple: It is the same mechanical

principle used by your kitchen scale. A spring with a defined elasticity is

elongated or compressed due to the weight of the object to be measured. The

compression ∆z of the spring (with spring constant cz ) is a direct measure of

the force F exerted, which in the regime of elastic deformation obeys Hooke’s

law:

F = cz × ∆z. (1.1)

The only difference to the kitchen scale is the sensitivity of the measure-

ment. The “spring" is a bendable cantilever with a stiffness of 0.01 N/m to

10 N/m. Since interatomic forces are in the range of some nN, the cantilever

will be deflected by 0.01 nm to 100 nm. Consequently, the precise detection of

the cantilever bending is the key feature of an atomic force microscope. If a

sufficiently sharp tip is directly attached to the cantilever, we could then mea-

sure the interacting forces between the last atoms of the tip and the sample

through the bending of the cantilever.

In 1986 Binnig, Quate and Gerber have presented exactly this concept for the

first atomic force microscope [11]. They measured the deflection of a cantilever

with sub-Ångström precision by a scanning tunneling microscope (STM) [12]

and used a gold foil as the spring. The tip was a piece of diamond glued to

this home-made cantilever (see Fig. 1.1). With this set-up they imaged sample

topographies down to the nanometer scale.

a) b)

Fig. 1.1 a) The basic concept of the first atomic force microscope build in 1986 by Binnig,

Quate, and Gerber. A sharp diamond tip glued to a gold foil scanned the surface, while the

bending of the cantilever was detected with scanning tunneling microscope. b) The ultimate

goal was to measure the force between the front atom of the tip and a specific sample atom.

(Reproduced from Ref. [11])

1.2 Principles of Atomic Force Microscopy 3

a) b)

Fig. 1.2 a) Scanning electron micrograph of a rectangular silicon cantilever with a width of

127 µm and a thickness of 4.4 µm. b) A different view of the same cantilever reveals that the

cross section of the cantilever is trapezoidal and the cantilever has two geometrical widths, a

smaller one on the tip side and a broader one on the backside side. Therefore, most manufac-

turers give a “mean width” which is (7.6 + 29.9) /2 = 18.75 µm for the cantilever shown. The

trapezoidal shape of the cantilever is caused by the anisotropic etching of the silicon during

the microfabrication of the cantilever. (Images courtesy of Boris Anczykowski, nanoAnalytics

GmbH; used with kind permission.)

1.2.2

Current Experimental Set-Ups

During the last years the experimental set-up has been modified while atomic

force microscopy became a widespread research tool. Twenty years after its

invention commercial atomic force microscopes are offered from various man-

ufactures. Although most of these instruments are designed for specific appli-

cations and environments, they are typically based on the following type of

sensors, detection methods, and scanning principles.

mostly from silicon and silicon nitride as rectangular or V-shaped cantilevers.

Spring constants and resonance frequencies of cantilevers depend on the ac-

tual mode of operation. For contact AFM measurements they are about 0.01 to

1 N/m and 5 to 100 kHz, respectively. In a typical force microscope, cantilever

deflections in the range from 0.1 Å to a few micrometers are measured. This

corresponds to a force sensitivity ranging from 10−13 N to 10−5 N.

Figure 1.2) shows two scanning electron microscope (SEM) images of a typ-

ical rectangular silicon cantilever. Using this imaging technique the length (l),

width (w), and thickness (t) can be precisely measured. The spring constant cz

of the cantilever can then be determined from these values [64]

3

w t

cz = ESi (1.2)

4 l

4 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

where ESi = 1.69 × 101 1 N/m2 is the Youngs’s modulus. Typical dimen-

sions of silicon cantilevers are as follows: lengths of 100–300 µm, widths of

10–30 µm, and thicknesses of 0.3–5 µm.

The torsion of the cantilever due to lateral forces between tip and surface

depends also on the height of the tip h. The torsional spring constant can be

calculated from [64]

G wt3

ctor = (1.3)

3 lh2

Since the dimensions of cantilevers given by the manufacturer are only av-

erage values, high accuracy calibration of the spring constant requires the

measurement of length, width, and thickness for each individual cantilever.

The length and the width can be measured with sufficient accuracy using an

optical microscope, but the thickness requires high resolution techniques like

scanning electron microscopy. In order to avoid this time and cost consuming

measurement one can determine the cantilever thickness from its eigenfre-

quency in normal direction [10, 59, 64]

√ r

4 3 ρSi 2

t= l f0 (1.4)

0.5968612π ESi

| {z }

≈ 7.23×10−4 s/m

The formulas presented above are only valid for rectangular cantilevers.

Equations for V-shaped cantilevers can be found in Refs. [70, 78].

measure the minute bendings have been improved. Today, commercial AFMs

use the so-called laser beam deflection scheme shown in Fig. 1.3. The bending

and torsion of cantilevers can be detected by a laser beam reflected from their

backside [3, 65], while the reflected laser spot is detected with a sectioned

photo-diode. The different parts are read out separately. Usually a four-

quadrant diode is used, to detect the normal as well as the torsional move-

ments of the cantilever. With the cantilever at equilibrium the spot is adjusted

such that the upper and the lower sections show the same intensity. If the can-

tilever bends up or down, the spot moves, and the difference signal between

upper and lower section is a measure of the bending.

The sensitivity can be improved by interferometer systems adapted by sev-

eral research groups (see, e.g., [5,43,69,77]). It is also possible to use cantilevers

with integrated deflection sensors based on piezoresistive films [55, 99, 102].

Since no optical parts are needed in the experimental set-up of an atomic force

1.2 Principles of Atomic Force Microscopy 5

photo−diode

mirror laser

torsion AB

deflection CD

set−point

cantilever

error signal

PID

scanner z

y

x

z−position

Fig. 1.3 Principle of an atomic force microscope working with the laser beam deflection

method. Deflection (normal force) and torsion (friction) of the cantilever are measured simul-

taneously by measuring the lateral and vertical deflection of a laser beam while the sample is

scanned in the x-y-plane. The laser beam deflection is determined using a 4-quadrant photo

diode: If A, B, C and D are proportional to the intensity of the incident light of the correspond-

ing quadrant, the signal (A+B)-(C+D) is a measure for the deflection and (A+C)-(B+D) a mea-

sure for the torsion of the cantilever. A schematic of the feedback system is shown by solid

lines. The actual deflection signal of the photo diode is compared with the set-point choosen

by the experimentalist. The resulting error signal is fed into the PID controller which moves the

z-position of the scanner in order to minimize the deflection signal.

slow scan direction

Fig. 1.4 Schematic of the scan process. The cantilever scans the sample surface sys-

tematically in x- and y-direction. Typical scan sizes ranging from less than 1 nm × 1 nm to

150 nm× 150 nm can be used.

microscope with these cantilevers the design can be very compact [88]. A

thorough comparison of different detection methods can be found in Ref. [10].

Scanning and Feedback System During scanning of the surface the deflec-

tion of the cantilever is kept constant by a feedback system, which controls

the vertical movement of the scanner. A schematic of the feedback system is

drawn in Fig. 1.3. It works as follows: The current signal of the photo-diode is

compared with a preset value. The feedback system including a proportional,

6 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

topography

error signal

Fig. 1.5 A simple lab class experiment demonstrating the scanning process of an atomic force

microscopy in contact mode. The cantilever is scanned over a simple test grid made of silicon.

The feedback keeps the deflection (and therefore the force) constant and the z-position of the

scanner is interpreted as the topography of the sample. The resulting map is plotted as a gray

scale image (lighter areas correspond to higher topography, upper left graph). The simulta-

neously plotted error signal (lower graphs) demonstrates that the feedback fails to keep the

deflection constant at the step edges.

integral, and differential (PID) controller varies the z-movement of the scan-

ner to minimize the difference. As a consequence the tip-sample force is kept

practically constant for an optimal set-up of the PID parameters.

While the cantilever is moving relative to the sample in the x-y-plane of

the surface by a piezoelectric scanner (see Fig. 1.4, the current z-position of

the scanner is recorded as a function of the lateral x-y-position with (ideally)

sub-Ångström precision. The obtained data represents a map of equal forces,

which is analyzed and visualized by computer processing.

Figure 1.5 shows a simple lab class experiment: imaging of a test grid. The

comparison between the topography and the error signal shows that the PID

controller needs some time at the step edges to correct the actual deflection

error.

1.2 Principles of Atomic Force Microscopy 7

1.2.3

Tip-Sample Forces in Atomic Force Microscopy

tected with an atomic force microscope. The obtained contrast depends on the

operational mode and the actual tip-sample interactions. Before discussing

details of the operational modes of atomic force microscopy we first specify

the most important tip-sample interactions.

Figure 1.6 shows the typical shape of the interaction force curve the tip

senses during an approach towards the sample surface. Upon approach of

the tip towards the sample, the negative attractive forces, representing e.g.

van der Waals or electrostatic interaction forces, increase until a minimum is

reached. This turnaround point is due to the onset of repulsive forces, caused

by Pauli repulsion, which will start to dominate upon further approach. Even-

tually, the tip is pushed into the sample surface and elastic deformation will

occur.

In general, the effective tip-sample interaction force is a sum of different

force contributions. They can be roughly divided into attractive and repulsive

components. The most important forces are summarized in the following.

Van der Waals forces are caused by fluctuating induced electric dipoles in

atoms and molecules. The distance dependence of this force for two distinct

molecules follows 1/z7 . For simplicity solid bodies are often assumed to con-

sist of many independent non-interacting molecules and the van der Waals

forces of these bodies are obtained by simple summation. For example, for a

sphere over a flat surface the van der Waals force is given by

AH R

FvdW (z) = − , (1.5)

6z2

where R is the radius of the sphere and A H is the Hamaker constant which

is typically in the range of ≈ 0.1 aJ [40]. This geometry is often used to ap-

proximate the van der Waals forces between tip and sample. Due to the 1/z2

dependency van der Waals forces are considered long-range forces compared

to other forces occuring in atomic force microscopy.

condense at the sample surface (and also on the tip) and cause the occurrence

of an adsorption layer. Consequently, the AFM tip penetrates through this

layer when approaching the sample surface. At the tip-sample contact, a wa-

ter meniscus is formed which causes a very strong attractive force [93]. For

soft samples such forces often lead to unwanted deformations of the surface.

However, this effect can be circumvented by measuring directly in liquids. Al-

8 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

a glove box with dry gases or in vacuum.

Pauli or Ionic Repulsion are the most important forces in conventional con-

tact mode AFM. The Pauli exclusion principle forbids that the charge clouds of

two electrons showing the same quantum numbers can have some significant

overlap; first, the energy of one of the electrons has to be increased. This yields

a repulsive force. Additionally, overlap of the charge clouds of electrons can

cause an insufficient screening of the nuclear charge, leading to ionic repul-

sion of Coulombic nature. The Pauli and the ionic repulsion are nearly hard

wall potentials. Thus, for tip and sample in intimate contact, most of the (re-

pulsive) interaction is carried by the atoms directly at the interface. The Pauli

repulsion is of purely quantum mechanical origin and semi-empirical poten-

tials are mostly used to allow an easy and fast calculation. A well-known

model is the Lennard-Jones potential, which combines short range repulsive

interactions with long range attractive van der Waals interactions:

r 6

r0 12

VLJ (z) = E0 −2 0 (1.6)

z z

where E0 is the bonding energy and r0 the equilibrium distance. In this case

the repulsion is described by an inverse power law with n = 12. The term

with n = 6 describes the attractive van der Waals potential between two

atoms/molecules.

Elastic forces and deformations can occur, if the tip is in contact with the

sample. Since this deformation affects the effective contact area the knowledge

about the elastic forces and the corresponding deformation mechanics of the

contact is an important issue in atomic force microscopy. The repulsive forces

occurring during elastic indentation of a sphere into a flat surface were already

analyzed in 1881 by H. Hertz (see, e.g., Refs. [41, 49])

4 ∗√

FHertz (z) = E R(z0 − z)3/2 for z ≤ z0 , (1.7)

3

where the effective elastic modulus E∗

1 (1 − µ2t ) (1 − µ2s )

= + (1.8)

E∗ Et Es

depends on the Young’s moduli Et,s and the Poisson ratios µt,s of tip and sur-

face, respectively. R is the tip radius and z0 is the point of contact.

However, this model does not include adhesion forces, which have to be

considered at the nanometer scale. Two extreme cases were analyzed by John-

son et al. [42] and Derjaguin et al. [16]. The model of Johnson, Kendall, and

1.2 Principles of Atomic Force Microscopy 9

6

tip-sample force (nN)

4 van-der-Waals force

2 Hertz model

-2

-4

-1.5 -1.0 -0.5 0.0 z 0 0.5 1.0 1.5 2.0

Fig. 1.6 Tip-sample model force after the DMT-M model for air Eq. (1.9) using the parameters

given in the text. The dashed line marks the position z0 where the tip touches the surface.

Roberts (JKR model) considers only the adhesion forces inside the contact area,

whereas the model of Derjaguin, Muller, and Toporov (DMT model) includes

only the adhesion outside the contact area. Various models analyzing the con-

tact mechanics in the intermediate regime were suggested by other authors

(see, e.g., Schwarz [87] for a recent overview).

However, in many practical cases it is sufficient to assume that the geo-

metrical shape of tip and sample does not change until contact has been es-

tablished at z = z0 and that afterwards, the tip-sample forces are given by

the DMT-M theory, denoting Maugis’ approximation to the earlier Derjaguin-

Muller-Toporov model [87]. In this approach, an offset FvdW (z0 ) is added to

the well-known Hertz model, which accounts for the adhesion force between

tip and sample surface. Therefore, the DMT-M model is often also referred to

as Hertz-plus-offset model [87]. The resulting overall force law is given by

− A H R for z ≥ z0 ,

2

FDMT−M (z) = 4 6z √ (1.9)

3 E∗ R(z0 − z)3/2 − A H2R for z < z0 .

6z 0

Figure 1.6 displays the resulting tip-sample force curve for the DMT-M

model. The following parameters were used, representing typical values for

AFM measurements under ambient conditions: AH = 0.2 aJ, R = 10 nm,

z0 = 0.3 nm, µt = µs = 0.3, Et = 130 GPa, and Es = 1 GPa.

Frictional forces are counteracting the movement of the tip during the scan

process. These forces dissipate the kinetic energy of the moving tip-sample

contact into the surface or tip material. This can be due to permanent changes

in the surface itself, by scratching or indenting, or also by the excitation of

lattice vibration (i.e. phonons) in the material.

Chemical binding forces arise from the overlap of molecular orbitals, due to

specific bonding states between the tip and the surface molecules. These force

10 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

v wear liquid film

F meniscus

force frictional

spring electrostatic trace of

forces forces the tip

step

tip + +

+

Pauli repulsion forces plastic deformation forces

adhesion

van der Waals forces

Fig. 1.7 Summary of the forces relevant in atomic force microscopy. (Image courtesy by Udo

D. Schwarz, Yale University; used with kind permission)

imaging of surfaces. Since these forces are also specific to the chemical identity

of the molecules, it is conceivable to identify the chemical character of the

surface atoms with AFM scans.

attractive or repulsive. They are usually measured when the tip is not in con-

tact with the surface (i.e. “non-contact”-mode). For magnetic forces, magnetic

materials have to be used for tip or tip coating. Well-defined electrical poten-

tials between tip and sample are necessary for the measurement of electro-

static forces.

More detailed information on intermolecular and surface forces relevant for

atomic force microscopy measurements can be found in the monographs of

Israelachvili [40] and Sarid [83]. Figure 1.7 summarizes the most important

forces. In principle every type of force can be measured with an atomic force

microscope. Nonetheless, the actual sensitivity to a specific force depends on

the mode of operation. Therefore, we introduce the most important modes in

the next section.

1.3

Modes of Operation

Here we concentrate on the two most important ones which are widely used

to image sample surfaces down to the atomic scale.

1.3 Modes of Operation 11

DPPC (L-α-dipalmitoyl-phophatidycholine) film adsorbed on mica. The image is color-coded,

i.e., dark areas represent the mica substrate and light areas the DPPC film. b) The simultane-

ously recorded friction image shows lower friction on the film (dark areas) as on the substrate

(light areas). The graphs at the bottom represent single scan lines obtained at the positions

marked by a dark line in the images at the top.

1.3.1

Static or Contact Mode

The contact mode is historically the oldest one. This widespread technique can

be used to obtain nanometer resolution images on a wide variety of surfaces.

Furthermore, it has the advantage that not only the deflection, but also the

torsion of the cantilever can be measured. As shown by Mate et al. [62] the

lateral force can be directly correlated to the friction between tip and sample,

thus extending atomic force microscopy to friction force microscopy (FFM).

Figure 1.8a) and b) present typical applications of an atomic force micro-

scope driven in contact-mode. The images represent a measurement of a

DPPC film adsorbed on a mica substrate. The lateral force was simultaneously

recorded with the topography and shows a contrast between the DPPC film

and the substrate. This effect can be attributed to the different frictional forces

on DPPC and the mica substrate and is frequently used to obtain a chemical

contrast on flat surfaces [63, 72].

12 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

repulsive

snap-in

retraction

0

approach

attractive

snap-out

Fadh

piezo movement

Fig. 1.9 A schematic of a typical force-vs-distance-curve obtained in static mode. The can-

tilever is approached towards the sample surface. Due to strong attractive forces it “jumps”

(snap-in) towards the sample surface at a specific position. During retraction the tip is strongly

attracted by the surface and the snap-out point is considerably behind the snap-in point. This

results in a hysteresis between approach and retraction.

However, we have neglected one important issue for the operation of the AFM

so far: The mechanical stability of the measurement. In static AFM the tip is

slowly approached towards the surface and the attractive forces between tip

and sample have to be counteracted by the restoring force of the cantilever.

This fails, however, if the force gradient of the tip-sample forces is larger than

the spring constant of the cantilever. Mathematically speaking: an instability

occurs if

∂ Fts (z)

cz < . (1.10)

∂z

In this case the attractive forces can no longer be sustained by the cantilever

and the tip “jumps” towards the sample surface [14].

This effect strongly influences static mode AFM measurements, as exem-

plified by a typical force-vs-distance curve shown in Fig. 1.9. Here, the force

acting on the tip recorded during an approach and retraction movement of the

cantilever is depicted. Upon approach of the cantilever towards the sample,

the attractive forces acting on the tip bend the cantilever towards the sam-

ple surface. At a specific point close to the sample surface these forces can

be no longer sustained by the cantilever spring and the tip “jumps” towards

the sample surface. Now tip and sample are in direct mechanical contact. A

further approach towards the sample surface pushes the tip into the sample.

1.3 Modes of Operation 13

Since the spring constant of the cantilever usually is much softer than the elas-

ticity of the sample the bending of the cantilever increases almost linearly.

If the cantilever is now retracted from the surface the tip stays in contact

with the sample, because it is strongly attracted by the sample due to adhesive

forces and the force Fadh is necessary to disconnect the tip from the surface.

The snap-out point is always at a larger distance from the surface than the

snap-in, which results in a hysteresis between approach and retraction of the

cantilever. This phenomenon of mechanical instability is often referred to as

the jump-to-contact. Unfortunately, this sudden jump can lead to undesired

changes of the tip and/or sample.

1.3.2

Dynamic Modes

Despite the success of contact mode AFM, the resolution was found to be

limited in many cases (in particular for soft samples) by lateral forces acting

between tip and sample. In order to avoid this effect, the cantilever can be

oscillated in vertical direction near the sample surface. AFM imaging with

vibrating cantilever is often denoted as dynamic force microscopy (DFM).

The historically oldest scheme of cantilever excitation in DFM imaging is

the external driving of the cantilever at a fixed excitation frequency exactly at

or very close to the cantilever’s first resonance [61, 75, 103]. For this driving

mechanism, different detection schemes measuring either the change of the

oscillation amplitude or the phase shift were proposed. Over the years, the

amplitude modulation (AM) or “tapping” mode, where the oscillation ampli-

tude is used as a measure of the tip-sample distance, has developed into the

most widespread technique for imaging under ambient conditions and liq-

uids.

In vacuum external oscillation of the cantilever has a disadvantage. Stan-

dard AFM cantilevers made from silicon exhibit very high Q values in vac-

uum, which results in very long response times of the system. Therefore, Al-

brecht et al. [2] introduced in 1991 the frequency modulation (FM) mode, which

works well for high-Q systems and consequently has developed into the dom-

inating driving scheme for high resolution DFM experiments in ultrahigh vac-

uum (UHV) [21,25,36,68]. In contrast to the AM mode, this approach features

a so-called self-driven oscillator [32, 33], which, in a closed-loop setup (“active

feedback”), uses the cantilever deflection itself as the driving signal, thus en-

suring that the cantilever instantaneously adapts to changes in the resonance

frequency. In the following section we discuss these two driving mechanisms

in more detail.

14 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

electronics with

laser photo diode lock−in amplifier

phase

amplitude

amplitude

z

piezo

D+2A aexc

d function

generator

D sample

PID

setpoint

x−y−z−scanner

Fig. 1.10 Set-up of a dynamic force microscope operated in the AM or tapping mode. A laser

beam is deflected by the back side of the cantilever, and its deflection is detected by a split

photo-diode. The cantilever vibration is caused by an external frequency generator driving an

excitation piezo. A lock-in amplifier is used to compare the cantilever driving with its oscillation.

The amplitude signal is held constant by a feedback loop controlling the cantilever sample

distance.

1.3.3

Amplitude-Modulation or Tapping Mode

As an alternative to the contact-mode the cantilever can be excited to vibrate

near its resonant frequency close to the sample surface. Under the influence

of tip-sample forces the resonant frequency (and consequently also amplitude

and phase) of the cantilever will change and serve as the measurement pa-

rameters. This is called the dynamic-mode. If the tip is approached towards

the surface, the oscillation parameters amplitude and phase are influenced by

the tip-surface interaction, and can therefore be used as feedback channels. A

certain set-point for e.g. the amplitude is given, and the feedback loop will

adjust the tip-sample distance such that the amplitude remains constant. The

controller parameter is recorded as a function of the lateral position of the tip

with respect to the sample and the scanned image essentially represents the

surface topography.

The technical realization of dynamic mode AFMs is based on the same key

components as a static AFM set-up. A sketch of the experimental setup of an

atomic force microscope driven in amplitude modulation mode is shown in

Fig. 1.10.

1.3 Modes of Operation 15

phophatidycholine) film adsorbed on mica. b) The phase contrast is directly related to the

topography, i.e., the phase is different between substrate and DPPC film.

The deflection of the cantilever is typically measured with the laser beam

deflection method as indicated [3, 65], but other displacement sensors such as

interferometric sensors [61, 69, 77, 86] have been applied as well. During op-

eration in conventional tapping mode, the cantilever is driven at a fixed fre-

quency with a constant excitation amplitude from an external function gen-

erator, while the resulting oscillation amplitude and/or the phase shift are

detected by a lock-in amplifier. The function generator supplies not only the

signal for the dither piezo; its signal serves simultaneously as a reference for

the lock-in amplifier.

This set-up can be operated in air and in liquids. A typical image obtained

with this experimental set-up in ambient conditions is shown in Fig. 1.11. For

a direct comparison with the static mode the sample is also DPPC adsorbed

on a mica substrate. While in contact-mode the frictional forces are measured

simultaneously with the topography, in dynamic-mode the phase between ex-

citation and oscillation is acquired as an additional channel. The phase im-

age gives information about the different material properties of DPPC and the

mica substrate. It can been shown, that the phase signal is closely related to

the energy dissipated in the tip-sample contact [15, 20, 98].

16 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

1.0

height (nm)

0.5

0.0

0 100 200 300 400 500

scan-position (nm)

Fig. 1.12 Topography of DNA adsorbed on mica imaged in buffer solution by tapping mode

AFM. The right graph shows a single scan line obtained at the position marked by an arrow in

the left image.

Due to its technical relevance the investigation of polymers has been the

focus of many studies (see, e.g., Ref. [60] for a recent review). High resolution

imaging has been extensively performed in the area of material science. Using

specific tips with additionally grown sharp spikes Klinov et al. [44] obtained

true molecular resolution on a polydiacetylene crystal.

Imaging in liquids opens up the avenue for the investigation of biologi-

cal samples in their natural environment. For example Möller et al. [66] have

obtained high-resolution images of the topography of hexagonally packed in-

termediate (HPI) layer of Deinococcus radiodurans with tapping-mode AFM.

A typical example for the imaging of DNA in liquid solution is shown in

Fig. 1.12.

Based on the above description of the experimental setup, we can formulate

the basic equation of motion describing the cantilever dynamics of amplitude-

modulation atomic force microscopy:

2π f 0 m

mz̈(t) + ż(t) + cz (z(t) − d)+ =

Q0

a cz cos(2π f d t) + Fts [z(t), ż(t)] . (1.11)

|d {z } | {z }

external driving force tip-sample force

p

Here, z(t) is the position of the tip at the time t; cz , m, and f 0 = (cz /m)/(2π )

are the spring constant, the effective mass, and the eigenfrequency of the can-

tilever, respectively. As a small simplification, we assume that the quality

factor Q0 combines the intrinsic damping of the cantilever and all influences

from surrounding media, such as air or liquid, if present, in a single overall

1.3 Modes of Operation 17

value. The equilibrium position of the tip is denoted as d. The first term on

the right hand side of the equation represents the external driving force of

the cantilever by the frequency generator. It is modulated with the constant

excitation amplitude ad at a fixed frequency f d . The (non-linear) tip-sample

interaction force Fts is introduced by the second term.

Before discussing the solutions of this equation, we would like to add some

words of caution regarding the universality of the equation of motion and the

various solutions discussed below. Eq. (1.11) disregards two effects, which

might become important under certain circumstances. First, we describe the

cantilever by a spring-mass-model and neglect in this way higher modes of the

cantilever. This is justified in most cases as the first eigenfrequency is by far

dominant in typical AM-AFM experiments (see, for example, [15,52,76,90,91]).

Secondly, we assume in our model equation of motion that the dither piezo

applies a sinusoidal force to the spring, but do not consider that the movement

of the dither piezo simultaneously also changes the effective position of the tip

at the cantilever end by aexc (t) = ad cos(2π f d t) [52–54]. This effect becomes

important when ad is in the range of the cantilever oscillation amplitude.

In a first step we assume that the cantilever vibrates far away from the sam-

ple surface. Consequently, we can neglect tip-sample forces (Fts ≡ 0), resulting

in the well-known equation of motion of a driven-damped harmonic oscilla-

tor.

After some time the external driving amplitude forces the cantilever to os-

cillate exactly at the driving frequency f d . Therefore, the steady-state solution

is given by the ansatz

where φ is the phase difference between the excitation and the oscillation of

the cantilever. With this we obtain two functions for the amplitude and phase

curves:

ad

A = s 2 , (1.13a)

2 2

fd 1 fd

1− +

f 02 Q0 f 0

1 fd / f0

tan φ = . (1.13b)

Q0 1 − f d 2 / f 0 2

The features of such an oscillator are well-known from introductory physics

courses.

If the cantilever is brought closer towards the sample surface the tip senses

the tip-sample interaction force Fts , which changes the oscillation behavior of

the cantilever. However, since the mathematical form of realistic tip-sample

forces is highly non-linear this fact complicates the analytical solution of the

18 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

equation of motion Eq. (1.11). For the analysis of dynamic force microscopy

experiments we need to focus on steady-state solutions of the equation of mo-

tion with sinusoidal cantilever oscillation. Therefore, it is advantageous to

expand the tip-sample force into a Fourier series

Z 1/ f

d

Fts [z(t), ż(t)] ≈ f d Fts [z(t), ż(t)]dt

0

Z 1/ f

d

+ 2 fd Fts [z(t), ż(t)] cos(2π f d t + φ)dt × cos(2π f d t + φ)

0

Z 1/ f

d

+ 2 fd Fts [z(t), ż(t)] sin(2π f d t + φ)dt × sin(2π f d t + φ)

0

+... ,

(1.14)

where z(t) is given by Eq. (1.12).

The first term in the Fourier series reflects the averaged tip-sample force

over one full oscillation cycle, which shifts the equilibrium point of the oscil-

lation by a small offset ∆d from d to d0 . Actual values for ∆d, however, are

very small. For typical amplitudes used in AM-AFM in air (some nm to some

tens of nm), the averaged tip-sample force is in the range of some pN. The

resulting offset ∆d is less than 1 pm for typical sets of parameters. Since this is

well beyond the resolution limit of an AM-AFM experiment in air, we neglect

this effect in the following and assume d ≈ d0 and D = d − A.

For further analysis, we now insert the first harmonics of the Fourier series

Eq. (1.14) into the equation of motion Eq. (1.11), obtaining two coupled equa-

tions [31, 81]

f 02 − f d2 ad

= I+ (d, A) + cos φ , (1.15a)

f 02 A

1 fd ad

− = I− (d, A) + sin φ , (1.15b)

Q0 f 0 A

where the following integrals have been defined:

2 f d 1/ f d

Z

I+ (d, A) = Fts [z(t), ż(t)] cos(2π f d t + φ)dt

cz A 0

1

Z d+ A

z−d

= ( F + F↑ ) p dz , (1.16a)

πcz A2 d− A ↓ A2 − ( z − d )2

2 f d 1/ f d

Z

I− (d, A) = Fts [z(t), ż(t)] sin(2π f d t + φ)dt

cz A 0

1

Z d+ A

= ( F − F↑ )dz

πcz A2 d− A ↓

1

= ∆E(d, A) . (1.16b)

πcz A2

1.3 Modes of Operation 19

Both integrals are functions of the actual oscillation amplitude A and cantilever-

sample distance d. Furthermore, they depend on the sum and the difference

of the tip-sample forces during approach (F↓ ) and retraction (F↑ ) manifested

by the labels “+" and “−" for easy distinction. The integral I+ is a weighted

average of the tip-sample forces (F↓ + F↑ ). On the other hand, the integral I−

is directly connected to ∆E, which reflects the energy dissipated during an

individual oscillation cycle. Consequently, this integral vanishes for purely

conservative tip-sample forces, where F↓ and F↑ are identical. A more detailed

discussion of these integrals can be found in Refs. [19] and [80].

By combining Eq. (1.13b) and (1.16b) we get a direct correlation between the

phase and the energy dissipation 1

A fd Q0 ∆E

sin φ = − + . (1.17)

A0 f 0 πcz A0 A

This relationship can be also obtained from the conservation of energy princi-

ple [15, 20, 98].

Equations (1.15) can be used to calculate the resonance curves of a dynamic

force microscope including tip-sample forces. The results are

ad

A = s 2 , (1.18a)

2 2

fd 1 fd

1− − I+ (d, A) + + I− (d, A)

f 02 Q0 f 0

1 fd

Q0 f 0 + I− (d, A)

tan φ = 2

. (1.18b)

fd

1− − I+ (d, A)

f 02

Equation (1.18a) describes the shape of the resonance curve, but it is an im-

plicit function of the oscillation amplitude A and cannot be plotted directly.

Figure 1.13 contrasts the solution of this equation (solid lines) with nu-

merical solution (symbols). As pointed out by various authors (see, e.g.,

Refs. [1,27,45,52,71,82,84,101]), the amplitude vs. frequency curves are multi-

valued within certain parameter ranges. Moreover, as the gradient of the an-

alytical curve increases to infinity at specific positions, some branches are un-

stable. The resulting instabilities are reflected by the “jumps” in the simulated

curves (marked by arrows in Fig. 1.13), where only stable oscillation states are

obtained. Obviously, they are different for increasing and decreasing driv-

ing frequency. This is a well-known effect frequently observed in non-linear

oscillators (see, e.g., Refs. [48] and [39]).

In amplitude-modulation atomic force microscopy, the cantilever might be

oscillated at any frequency around the resonance peak. Here, we restrict our-

1) The “-”-sign on the right side of the equation is due to our definition

of the phase φ in Eq. (1.12).

20 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

11

tapping (Q = 300)

amplitude (nm)

10

9

8

7

6

299.2 299.6 300.0 300.4 300.8

Fig. 1.13 Resonance curves for tapping mode operation if the cantilever oscillates near the

sample surface with d = 8.5 nm and A0 = 10 nm, thereby experiencing the model force

field given by Eq. (1.9). The solid lines represent the analytical result of Eq. (1.18a), while the

symbols are obtained from the numerical solution of the equation of motion Eq. (1.11). The

dashed lines reflect the resonance curves without tip-sample force and are shown purely for

comparison. The resonance curve exhibits instabilities (“jumps”) during a frequency sweep.

These jumps take place at different positions (marked by arrows) depending on whether the

driving frequency is increased or decreased.

selves to the situation where the driving frequency is set exactly to the eigen-

frequency of the cantilever ( f d = f 0 ). With this choice, which is also very com-

mon in actual DFM experiments, we have defined imaging conditions leading

to handy formulas suitable for further analysis. A discussion on the alterna-

tive cases of driving the cantilever slightly above or slightly below resonance

can be found in the following Refs. [6, 29, 92]

From Eq. (1.18a), we obtain the following relationship between the free os-

cillation amplitude A0 , the actual amplitude A, and the equilibrium tip posi-

tion d:

q

A0 = A 1 + ( Q0 I+ [d, A])2 . (1.19)

For the derivation of this formula, we used the approximation that the maxi-

mal value of the free oscillation amplitude at resonance is given by A0 ≈ ad Q0 .

Solving this equation allows us to study amplitude vs. distance curves for

different effective Q-factors, as shown in Fig. 1.14 for a Q-factor of 300. As

observed before in the resonance curves displayed in Fig. 1.13 stable and un-

stable branches develop, which can be unambiguously identified by a com-

parison with numerical results (symbols).

Most noticeable, the tapping mode curve exhibits jumps between unsta-

ble branches, which occur at different locations for approach and retraction.

The resulting bistable regime then causes a hysteresis between approach and

retraction, which has been the focus of numerous experimental and theoreti-

cal studies (see, e.g., Refs. [6, 22, 29, 52, 74, 82, 97, 104]). As shown by various

authors, the instability in conventional AM-AFM divides the tip-sample in-

teraction into two regimes [52, 74, 82, 104]. Before the instability occurs the tip

1.3 Modes of Operation 21

10

bistable regime

8

amplitude (nm)

A(d)-curve

approach

2

retraction

0

0 2 4 6 8 10 12

distance d (nm)

Fig. 1.14 Amplitude vs. distance curve for conventional (“tapping mode”) AM-AFM for A0 =

10 nm, f 0 = 300 kHz, and a tip-sample interaction force as given in Fig. 1.6. The dashed lines

represent the analytical result, while the symbols are obtained from the numerical solution

of the equation of motion Eq. (1.11). The overall amplitudes decrease during an approach

towards the sample surface in both cases, but instabilities (indicated by arrows) occur.

of the tip-sample force. After jumping to the higher branch, however, the tip

senses also the repulsive part of the tip-sample interaction.

In Fig. 1.15 the oscillation amplitude is plotted as a function of the nearest

tip-sample distance. In addition, the bottom graph depicts the corresponding

tip-sample force (cf. Fig. 1.6). The origin of the nearest tip-sample position

D is defined by this force curve. Since both the amplitude curves and the tip-

sample force curve are plotted as a function of the nearest tip-sample position,

it is possible to identify the resulting maximum tip-sample interaction force

for a given oscillation amplitude.

A closer look at the A( D )-curves helps now identifying the different inter-

action regimes in AM-AFM. During the approach of the vibrating cantilever

towards the sample surface, this curve shows a discontinuity for the nearest

tip-sample position D (point of closest approach during an individual oscilla-

tion) between 0 and -1 nm. This gap corresponds to the bi-stability and the

resulting jumps in the amplitude vs. distance curve. After the jump from

the attractive to the repulsive regime has occurred, the amplitude decreases

continuously. The nearest tip-sample position, however, does not reduce ac-

cordingly, remaining roughly between -0.8 nm and -1.5 nm. As a result, larger

22 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

10

amplitude (nm)

4 approach

retraction

2

0

tip-sample force (nN)

6 tapping

regime

4

bistable

2 tapping

regime

0

-2

-4

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0

nearest tip-sample distance (nm)

Fig. 1.15 A comparison between the maximum tip-sample forces (tip-sample forces acting

at the point of closest tip-sample approach/nearest tip-sample position D ) experienced by

conventional “tapping mode” AM-AFM assuming the same parameters as in Fig. 1.14. The

upper graph shows the nearest tip-sample position D vs. the actual oscillation amplitude A

for tapping mode. The graph at the bottom reveals the force regimes sensed by the tip. The

maximal tip-sample forces in tapping mode are on the repulsive (tapping regime) as well as

attractive (bistable tapping regime) part of the tip-sample force curve.

which is important to keep in mind while adjusting imaging parameters in

tapping mode AM-AFM imaging. In contrast, once the repulsive regime has

been reached, the user’s ability to influence the tip-sample interaction strength

by modifying the set-point for A is limited, thus also limiting the possibilities

of improving the image quality.

For practical applications, it is reasonable to assume that the set-point of

the amplitude used for imaging has been set to a value between 90 % (= 9 nm)

and 10 % (= 1 nm) of the free oscillation amplitude. With this condition, we can

identify the accessible imaging regimes indicated by the horizontal (dashed)

lines and the corresponding vertical (dotted) lines. In tapping mode, two

imaging regimes are realized: the tapping regime (left) and the bistable tapping

regime (middle). The first can be accessed by any amplitude set-point between

9 nm and 1 nm and results in a maximum tip-sample forces well within the

repulsive regime. The second one, belonging to the bistable imaging state, is

only accessible during approach. The corresponding amplitude set-point is

1.3 Modes of Operation 23

between 9 nm and 8 nm. Imaging in this regime is possible with the limitation

that the oscillating cantilever might jump into the repulsive regime [74, 92].

In the above paragraphs, we have outlined the influence of the tip-sample in-

teraction on the cantilever oscillation, calculated the maximum tip-sample in-

teraction forces based on the assumption of a specific model force, and subse-

quently discussed possible routes for image optimization. However, in prac-

tical imaging, the tip-sample interaction is not a priori known. Several au-

thors [30, 51, 54, 89] have suggested solutions to this inversion problem. Here

we present an approach which is based on the analysis of the amplitude and

phase vs. distance curves which can be easily measured with most AM-AFM

set-ups.

Let us start by applying the transformation D = d − A to the integral I+

Eq. (1.16a), where D corresponds to the nearest tip-sample distance as defined

in Fig. 1.10. Next, we note that due to the cantilever oscillation, the current

method intrinsically recovers the values of the force that the tip experiences

at its lower turning point, where F↓ necessarily equals F↑ . We thus define

Fts = ( F↓ + F↑ )/2, and Eq. (1.16a) subsequently reads as

DZ+2A

2 z−D−A

I+ = Fts p dz . (1.20)

πcz A2 A2 − ( z − D − A )2

D

The amplitudes commonly used in AM-AFM are considerably larger than the

interaction range of the tip-sample force. Consequently, tip-sample forces in

the integration range between D + A and D + 2A are practicably insignif-

icant. For this so-called “large amplitude approximation”

p [17, 26], the last

term can be expanded at z → D to ( z − D − A ) / A 2 − ( z − D − A )2 ≈

p

− A/2(z − D ), resulting in

√ DZ+2A

2 Fts

I+ ≈ − √ dz . (1.21)

πcz A3/2 z−D

D

Introducing this equation into Eq. (1.15a) we obtain the following integral

equation:

" # DZ+2A

cz A3/2 ad cos(φ) f 02 − f d2 1 F

√ − 2

= √ ts dz . (1.22)

2 A f0 π z−D

| {z } D

κ

The left hand side of this equation contains only experimentally accessible

data, and we denote this term as κ. The benefit of these transformations is

24 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

that the integral equation can be inverted [17,48] and, as a final result, we find

DZ+2A

∂ κ (z)

Fts ( D ) = − √ dz . (1.23)

∂D z−D

D

It is now straightforward to recover the tip-sample force using Eq. (1.23) from

a “spectroscopy experiment”, i.e., an experiment where the amplitude and

the phase are continuously measured as a function of the actual tip-sample

distance D = d − A at a fixed location. With this input, one first calculates κ

as a function of D. In a second step, the tip-sample force is computed solving

the integral in Eq. (1.23) numerically.

Additional information about the tip-sample interaction can be obtained re-

memebering that the integral I− is directly connected to the energy dissipation

∆E. By simply combining Eqs. (1.15b) and (1.16b), we get

1 fd ad

∆E = + sin φ πcz A2 . (1.24)

Q0 f 0 A

The same result has been found earlier by Cleveland et al. [15] using the con-

servation of energy principle. However, in a further development of Cleve-

land’s work, we suggest to plot the energy dissipation as a function of the

nearest tip-sample distance D = d − A in order to have the same scaling as

for the tip-sample force.

An application of the method to experimental data obtained on a silicon

wafer is shown in Fig. 1.16 where we used only the data points before the jump

to reconstruct tip-sample force and energy dissipation. As a consequence the

experimental force curve shows only the attractive part of the force between

tip and sample with a minimum of -1.8 nN. This result is in agreement with

previous studies stating that the tip senses only attractive forces before the

jump [6, 22, 46].

1.3.4

Frequency-Modulation or Non-contact Mode in Vacuum

important to prepare clean sample surfaces free from unwanted adsorbates.

Therefore, these experiments are usually performed in ultra-high vacuum

with pressures below 1 × 10−10. As a consequence most dynamic force micro-

scope experiments in vacuum utilize the so-called frequency modulation (FM)

detection scheme introduced by Albrecht et al. [2]. In this mode the cantilever

is self-oscillated, in contrast to the AM- or tapping-mode discussed in Sec.

1.3.3. The FM-technique enables the imaging of single point defects on clean

sample surfaces in vacuum and its resolution is comparable with the scan-

ning tunneling microscope, while not restricted to conducting surfaces. In the

1.3 Modes of Operation 25

16

amplitude (nm) 0.0

12

-0.5

8

-1.0

4

-1.5

0 -2.0

0 5 10 15 20 25 30 4 5 6 7 8 9 10

(a) piezo movement (nm) (c) relative tip-sample position (nm)

12

phase (degree)

50

0 8

-50

4

-100

-150 0

0 5 10 15 20 25 30 4 5 6 7 8 9 10

60

)

-15

x10

40

1/2

20

?

kappa (Nm

-20

-40

2 4 6 8 10

Fig. 1.16 dynamic force spectroscopy experiment on a silicon wafer in air (parameters of the

cantilever: f d = f 0 = 328.61 kHz, c z = 33.45 N/m, Q0 = 537).a) A measurement of the os-

cillation amplitude as a function of the oscillation amplitude shows jumps at different positions

during approach and retraction. b) The jumps are also observed in the phase vs. distance

curves. c) Using the algorithm described in the text the tip-sample force can be reconstructed.

This curve is calculated from the approach data. Only the data points before the jump are

used for the reconstruction of the tip-sample force. d) The energy dissipation per oscillation

cycle can be easily obtained from equation Eq. (1.24). e) This graph shows the κ ( D )-values

computed from the amplitude and phase vs. distance curves plotted in (a) and (b). The jump in

these curves results also in a jump in the κ-curve.

years after the invention of the FM-technique the term non-contact atomic force

microscopy (NC-AFM) was established, because it is commonly believed that

a repulsive, destructive contact between tip and sample is prevented by this

technique. In this subsection we introduce the basic principles of the exper-

26 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

electronics with

laser photo diode frequency detector

amplitude

frequency

amplifier

phase

shifter θ0

frequency

z

piezo

D+2A aexc

d

D sample

PID

setpoint

x−y−z−scanner

Fig. 1.17 The schematic set-up of a dynamic force microscope using the frequency modula-

tion technique. This experimental set-up is often used in UHV. A significant feature is the pos-

itive feedback of the self-driven cantilever. The detector signal is amplified and phase shifted

before it is used to drive the piezo. The measured quantity is the frequency shift due to the

tip-sample interaction, which serves as the control signal for the cantilever-sample distance.

imental set-up, explain the origin and calculation of the detected frequency

shift and present applications of this mode.

In vacuum applications, the Q-factor of silicon cantilevers is in the range

of 10 000–30 000. High Q-factors, however, limit the acquisition time (band-

width) of a dynamic force microscopy, since the oscillation amplitude of the

cantilever needs a long time to adjust. This problem is avoided by the FM-

detection scheme based on the specific features of a self-driven oscillator.

The basic set-up of a dynamic force microscope utilizing this driving mech-

anism is schematically shown in Fig. 1.17. The movement of the cantilever is

measured with a displacement sensor. This signal is then fed back into an

amplifier with an automatic gain control (AGC) and is subsequently used to

excite the piezo oscillating the cantilever. The time delay between the excita-

tion signal and cantilever deflection is adjusted by a time (“phase”) shifter to

a value t0 = 1/(4 f 0 ) corresponding to ≈ 90◦ , since this ensures an oscillation

at resonance. Two different modes have been established: The constant ampli-

tude-mode [2], where the oscillation amplitude A is kept at a constant value

by the AGC, and the constant excitation mode [100], where the excitation am-

plitude is kept constant. In the following, however, we focus on the constant

amplitude mode.

1.3 Modes of Operation 27

The key feature of the described set-up is the positive feedback-loop which

oscillates the cantilever always at its resonance frequency f [33]. The reason

for this behavior is that the cantilever serves as the frequency determining ele-

ment. This is in contrast to an external driving of the cantilever by a frequency

generator, where the driving frequency f d is not necessarily the resonant fre-

quency of the cantilever.

If the cantilever oscillates near the sample surface, the tip-sample interac-

tion alters its resonant frequency, which is then different from the eigenfre-

quency f 0 of the free cantilever. The actual value of the resonant frequency

depends on the nearest tip-sample distance and the oscillation amplitude. The

measured quantity is the frequency shift ∆ f , which is defined as the difference

between both frequencies (∆ f := f − f 0 ). The detection method received its

name from the frequency demodulator (FM-detector). The cantilever driving

mechanism, however, is independent of this part of the set-up. Other set-ups

use a phase-locked-loop (PLL) to detect the frequency and to oscillate the can-

tilever exactly with the frequency measured by the PLL [56].

For imaging the frequency shift ∆ f is used to control the cantilever sample

distance. Thus, the frequency shift is constant and the acquired data repre-

sents planes of constant ∆ f , which can be related to the surface topography

in many cases. The recording of the frequency shift as a function of the tip-

sample distance or the oscillation amplitude can be used to determine the tip-

sample force with high resolution (see Sec. 1.3.4.5).

Before presenting experimental results obtained in vacuum we will analyze

the origin of the frequency shift. A good insight into the cantilever dynamics

is given by looking at the tip potential displayed in Fig. 1.18. If the cantilever

is far away from the sample surface, the tip moves in a symmetric parabolic

potential (dotted line), and its oscillation is harmonic. In such a case, the tip

motion is sinusoidal and the resonance frequency is given by the eigenfre-

quency f 0 of the cantilever. If, however, the cantilever approaches the sample

surface, the potential – which determines the tip oscillation – is modified to an

effective potential Veff (solid line) given by the sum of the parabolic potential

and the tip-sample interaction potential Vts (dashed line). This effective po-

tential differs from the original parabolic potential and shows an asymmetric

shape.

As a result of this modification of the tip potential the oscillation becomes

inharmonic, and the resonance frequency of the cantilever depends now on

the oscillation amplitude A. Since the effective potential experienced by the

tip changes also with the nearest distance D, the frequency shift is a functional

of both parameters (⇒ ∆ f := ∆ f ( D, A)).

28 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

V(z)

cantilever potential

tip−sample potential

effective potential

D + 2A

z

z min

D

Fig. 1.18 The frequency shift in dynamic force microscopy is caused by the tip-sample in-

teraction potential (dashed line), which alters the harmonic cantilever potential (dotted line).

Therefore, the tip moves in an anharmonic and asymmetric effective potential (solid line). zmin

is the minimum position of the effective potential.

curves for different oscillation amplitudes. The experiments were carried out

with an atomic force microscope designed for operation in ultrahigh vacuum

(UHV) and at low temperatures [5].

The obtained experimental frequency shift vs. distance curves show a be-

havior expected from the simple model explained above. All curves show a

similar overall shape, but differ in magnitude in dependence of the oscilla-

tion amplitude and the nearest tip-sample distance. During the approach of

the cantilever towards the sample surface, the frequency shift decreases and

reaches a minimum. With a further reduction of the nearest tip-sample dis-

tance, the frequency shift increases again and becomes positive. For smaller

oscillation amplitudes, the minimum of the ∆ f (z)-curves is deeper and the

slope after the minimum is steeper than for larger amplitudes, i.e., the overall

effect is larger for smaller amplitudes.

This can be explained by the simple potential model as well: A decrease

of the amplitude A for a fixed nearest distance D moves the minimum of the

effective potential closer to the sample surface. Therefore, the relative pertur-

bation of the harmonic cantilever potential increases, which increases also the

absolute value of the frequency shift.

As already described in the previous subsection it is a specific feature of the

FM-modulation technique that the cantilever is “self-driven” by a positive

1.3 Modes of Operation 29

0,8 0,10

]

-15

x 10

frequency shift [kHz]

0,4 0,05

1/2

norm. freq. shift [Nm

0,0 0,00

180 Å 180 Å

126 Å 126 Å

-0,4 90 Å -0,05 90 Å

72 Å 72 Å

54 Å 54 Å

-0,8 z0 5 -0,10

-10 -5 0 10 15 -10 -5 0 5 10 15

a) distance D [Å] b) distance D [Å]

8

tip-sample force [nN]

0

180 Å

126 Å

-4 90 Å

72 Å

Fad 54 Å

-8 z0 5

-10 -5 0 10 15

c) distance D [Å]

Fig. 1.19 a) Experimental frequency shift versus distance curves acquired with a silicon can-

tilever (c z = 38 N/m; f 0 = 171 kHz) and a graphite sample for different amplitudes (54–180 Å)

in UHV at low temperature ( T = 80 K). The curves are shifted along the x-axes to make

them comparable. b) Transformation of all frequency shift curves shown in a) to one univer-

sal curves using Eq. (1.32). The normalized frequency shift γ( D ) is nearly identical for all

amplitudes. c) The tip-sample force calculated with the experimental data shown in a) and b)

using the formula Eq. (1.33). The force Fts Eq. (1.34) is plotted by a dashed-dotted line. The

best fit using the force law Fc is displayed by a solid line. The border between “contact” and

“non-contact” force is marked by the position z0 .

motion is different from the case of the externally driven cantilever discussed

in Sec. 1.3.3. The external driving term has to be replaced in order to describe

the self-driving mechanism correctly. Therefore, the equation of motion is

30 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

given by

2π f 0 m∗

m∗ z̈(t) + ż(t) + cz (z(t) − d) + g cz (z(t − t0 ) − d) = Fts [z(t), ż(t)].

Q | {z }

driving

(1.25)

where z := z(t) represents the position of the tip at the time t; cz , m, and

Q are the spring constant, the effective mass, and the quality factor of the

cantilever, respectively. Fts = −(∂ Vts )/(∂z) is the tip-sample interaction force.

The last term on the left describes the active feedback of the system by the

amplification of the displacement signal by the gain factor g measured at the

retarded time t − t0 .

The frequency shift can be calculated from the above equation of motion

with the ansatz

assume that the cantilever oscillations are more or less sinusoidal and develop

the tip-sample force Fts into a Fourier-series like Eq. (1.14). This procedure

results in a set of two coupled trigonometric equations [32, 33]:

f 2 − f 02

g cos(2π f t0 ) = + I+ (1.27)

f 02

1 f

g sin(2π f t0 ) = + I− (1.28)

Q f0

where the two integrals I+ Eq. (1.16a) and I− Eq. (1.16b) were defined in accor-

dance to Sec. 1.3.3.2. These two coupled equations can be solved numerically,

if one is interested in the exact dependency of the tip-sample interaction force

Fts and the time delay t0 on the oscillation frequency f and the gain factor g.

Fortunately, a detailed analysis shows that the results of a FM-AFM exper-

iment are mainly determined by the tip-sample force and only very slightly

by the time delay, if t0 is set to an optimal value before approaching the tip

towards the sample surface. These values of the time delay are specific reso-

nance values corresponding to 90 deg (i.e. t0 = 1/4 f 0 ) and can be easily found

by minimizing the gain factor as a function of the time delay. Therefore, it can

be assumed that cos(2π f t0 ) ≈ 0 and sin(2π f t0 ) ≈ 1 and the two coupled

equations (1.27) and (1.28) can be decoupled. As a result we get an equation

for the frequency shift

1

Z d+ A

f0 z−d

∆f ∼

= − I+ = ( F↓ + F↑ ) p dz (1.29)

2 πcz A2 d− A A2 − ( z − d )2

1.3 Modes of Operation 31

1 f

∆E = g − πcz A2 . (1.30)

Q f0

As no assumptions were made about the specific force law of the tip-sample

interaction Fts , these equations are valid for any type of interaction as long as

the resulting cantilever oscillations remain nearly sinusoidal.

Since the amplitudes in FM-AFM are often considerably larger than the dis-

tance range of the tip-sample interaction, we can again make the “large am-

plitude approximation" [17,26] and introduce the approximation Eq. (1.21) for

the integral I− . This yields the formula

DZ+2A

1 f0 F (z)

∆f = √ √ ts dz (1.31)

2π cz A3/2 z−D

D

dent of the oscillation amplitude. The experimental parameters (cz , f 0 , and A)

appear as pre-factors. Consequently, it is possible to define the normalized fre-

quency shift [26]

cz A3/2

γ(z) := ∆ f ( z ). (1.32)

f0

amplitudes and cantilevers. The validity of Eq. (1.32) is nicely demonstrated

by the application of this equation to the frequency shift curves already pre-

sented in Fig. 1.19a). As shown in Fig. 1.19b) all curves obtained for different

amplitudes result into one universal γ-curve, which depends only on the ac-

tual tip-sample distance D.

The excitement about the NC-AFM technique in ultrahigh vaccum was driven

by the first results of Giessibl [23] who achieved to image the true atomic struc-

ture of the Si(111)-7 × 7-surface with this technique in 1995. In the same year

Sugawara et al. [95] observed the motion of single atomic defects on InP with

true atomic resolution. However, imaging on conducting or semi-conducting

surfaces is also possible with the scanning tunneling microscope (STM) and

these first NC-AFM images provided no new information on surface prop-

erties. The true potential of NC-AFM lies in the imaging of non-conducting

surface with atomic precision, which was first demonstrated by Bammerlin

et al. [8] on NaCl. A long-standing question about the surface reconstruction

of the technological relevant material Aluminium oxide could be answered

32 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

a) b)

Fig. 1.20 Imaging of a NiO(001) sample surface with a non-contact AFM. a) Surface step

and an atomic defect. The lateral distance between two atoms is 4.17 Å. b) A dopant atom is

imaged as a light protrusion about 0.1 Å higher as the other atoms. (Images courtesy of W.

Allers and S. Langkat, University of Hamburg; used with kind permission)

by Barth et al. [9], who imaged the atomic structure of the high temperature

phase of α-Al2 O3 (0001).

The high resolution capabilities of non-contact atomic force microscopy are

nicely demonstrated by the images shown in Fig. 1.20. Allers et al. [4] resolved

atomic steps and defects with atomic resolution on Nickel oxide. Today such

a resolution is routinely obtained by various research groups (for an overview

see, e.g., Refs. [21,25,64,67]). Recent efforts have also been concentrated on the

analysis of functional organic molecules, since in the field of nanoelectronics it

is anticipated that in particular organic molecules will play an important role

as the fundamental building blocks of nanoscale electronic device elements.

For example atomic resolution on the highly curved surface of a nanotube

[7] was achieved. The analysis of growth properties of thin films [13, 47, 58]

with respect to their electronic properties has been investigated, as well as the

intra-molecular contrast of individual molecules has been resolved [94], which

might be directly related to the internal charge density distribution inside the

molecules.

Since its invention in 1986 the atomic force microscope has been widely used

to study tip-sample interactions for various material combinations. Unfortu-

nately, in contact mode such investigations were often hindered close to the

sample surface by a “jump to contact” (see Sec. 1.3.1.1). In tapping mode on

the other hand the force analysis is limited due to the instabilities in the am-

plitude and phase vs. distance curves (see 1.3.3.3). Such problems, however,

are avoided using large oscillation amplitudes in the frequency modulation

technique.

1.3 Modes of Operation 33

In Sec. 1.3.4.3 we showed how the frequency shift can be calculated for a

given tip-sample interaction law. The inverse problem, however, is even more

interesting: How can the tip-sample interaction be determined from frequency shift

data? Various mathematical solutions to this question have been presented

by various authors [17, 18, 24, 28, 34, 38, 79] and have lead to the dynamic force

spectroscopy (DFS) technique, which is a direct extension of the FM-AFM mode.

Here we present the approach of Dürig [17], which is based on the inversion

of the integral Eq. (1.31) already presented in Sec. 1.3.4.3. It can be transformed

to

∞

√ cz A3/2 ∂ Z ∆ f (z)

Fts ( D ) = 2 √ dz, (1.33)

f 0 ∂D z−D

D

which allows a direct calculation of the tip-sample interaction force from the

frequency shift versus distance curves.

An application of this formula to the experimental frequency shift curves

already presented in Sec. 1.3.4.2 is shown in Fig. 1.19c). The obtained force

curves are nearly identical although obtained with different oscillation ampli-

tudes. Since the the tip-sample interactions can be measured with high resolu-

tion, dynamic force spectroscopy opens a direct way to compare experiments

with theoretical models and predictions.

Giessibl [26] suggested to describe the force between the tip and the sample

by a combination of a long-range (van der Waals) and a short-range (Lennard-

Jones) term (see Sec. 1.2.3). The long-range part describes the van der Waals

interaction of the tip, modelled as a sphere with a specific radius, with the sur-

face. The short-range Lennard-Jones term is a superposition of the attractive

van der Waals interaction of the last tip apex atom with the surface and the

Coulomb repulsion. For a tip with the radius R, this assumption results in the

tip-sample force:

AH R 12E0 r0 13 r0 7

Fnc (z) = − + − . (1.34)

6z2 r0 z z

Since this approach does not explicitly consider elastic contact forces between

tip and sample, we call this the “non-contact” force law in the following.

A fit of this equation to the experimental tip-sample force curve is shown

in Fig. 1.19 by a solid line; the obtained parameters are AH R = 2.4 · 10−27 Jm,

r0 = 3.4 Å, and E0 = 3 eV [37]. The regime on the right from the minimum

fits well to the experimental data, but the deep and wide minimum of the ex-

perimental curves cannot be described accurately with the non-contact force.

This is caused by the steep increase of the Lennard-Jones force in the repulsive

regime (⇒ Fts ∝ 1/r13 for z < r0 ).

The elastic contact behavior can be described with the assumption of the

above described DMT-M model (see Sec. 1.2.3) that the overall shape of tip

34 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

z

D+2A 0.4

z-direction / nm

0.3 -0.2 nN

D

y 0.2

0.1

-1.0 nN

0

0 0.2 0.4 0.6 0.8 1

x y-direction / nm

a) b)

Fig. 1.21 a) Principle of 3D-force spectroscopy. The cantilever oscillates near the sample sur-

face and measure the frequency shift in a x-y-z-box. The three dimensional surface shows the

topography of the sample (image size: 10 Å × 10 Å) obtained immediately before the recording

of the spectroscopy field. b) The reconstructed force field of NiO(001) shows atomic resolu-

tion. The data is taken along the line shown in a).

and sample changes only slightly until point contact is reached and that, af-

ter the formation of this point contact, the tip-sample forces are described by

the Hertz theory. A fit of the Hertz model to the experimental data is shown

in Fig. 1.19 by a solid line. The experimental force curves agree quite well

with the contact force law for distances D < z0 . This shows that the overall

behavior of the experimentally obtained force curves can be described by a

combination of long-range (van der Waals), short-range (Lennard-Jones), and

contact (Hertz/DMT) forces.

Since Eq. (1.31) was derived under the assumption that the resonance am-

plitude is considerably larger than the decay length of the tip-sample inter-

action, the same restriction applies for Eq. (1.33). However, using more ad-

vanced algorithms it is also possible to determine forces from dynamic force

spectroscopy experiments without the large amplitude restriction. The nu-

merical approach of Gotsmann et al. [28] as well as the semi-empirical meth-

ods by Dürig [18], Giessibl [24] and Sader and Jarvis [79] are applicable in all

regimes.

The resolution of dynamic force spectroscopy can be driven down to the

atomic scale. Lantz et al. [50] measured frequency shift vs. distance curves at

different lattice sites of the Si(111)-(7x7) surface. In this way they were able to

distinguish differences in the bonding forces between inequivalent adatoms

of the 7x7 surface reconstruction of Silicon.

The concept of dynamic force spectroscopy can be also extended to 3D-

force spectroscopy by mapping the complete force field above the sample

surface [35]. Figure 1.21a) shows a schematic of the measurement principle.

Frequency shift vs. distance curves are recorded on a matrix of points perpen-

dicular to the sample surface. Using Eq. (1.33) the complete 3-dimensional

1.3 Modes of Operation 35

0.04

0.02

0.00

-0.02

x-position (nm)

a) b)

Fig. 1.22 a) Figure shows a 3-dimensional representation of the interaction energy map deter-

mine from 3d force spectroscopy experiments on a NaCl(100) crystal surface. The red circular

depressions represent the local energy minima. b) This graph shows the potential energy pro-

file obtained from a) by collecting the energy minimum values along the x-axis. This curve thus

directly reveals the potential energy barrier of ∆Ebarrier = 48 meV separating the local energy

minima.

force field between tip and sample can be recovered with atomic resolution.

Figure 1.21b) shows a cut through the force field as a two-dimensional map.

The 3D-force technique has been applied also to a NaCl(100) surface, where

not only conservative but also the dissipative tip-sample interaction could be

measured in full space [85]. On one hand, the forces were measured in the

attractive as well as repulsive regime, allowing for the determination of the

local minima in the corresponding potential energy curves 1.22. This infor-

mation is directly related to the atomic energy barriers responsible for a mul-

titude of dynamic phenomena in surface science, like diffusion, faceting and

crystalline growth. The direct comparison of conservative with the simultane-

ously acquired dissipative processes furthermore allowed determining atomic

scale mechanical relaxation processes.

If the NC-AFM is capable of measuring forces between single atoms with

sub-nN precision, why should it not be possible to also exert forces with this

technique? In fact, the new and exciting field of nanomanipulation would be

driven to a whole new dimension, if defined forces can be reliably applied to

single atoms or molecules. In this respect, Loppacher et al. [57] achieved to

push on different parts of an isolated Cu-TBBP molecule, which is known to

possess four rotatable legs. They measured the force-distance curves while

one of the legs was pushed by the AFM tip and turned by 90 deg, thus be-

ing able to measure the energy which was dissipated during ’switching’ this

molecule between different conformational states. The manipulation of single

Silicon atoms with the NC-AFM was demonstrated by Oyabu et al. [73], who

removed single atoms from a Si(111)-7x7 surface with the AFM tip and could

subsequently deposit atoms from the tip on the surface again. This approach

was driven to its limits by Sugimoto et al. who manipulated single Sn-atoms

on the Ge(111)-c(2×8) semiconductor surface. By pushing single Sn-atoms

36 1 Nanoscale Imaging and Force Analysis with Atomic Force Microscopy

Fig. 1.23 Final topographic NC-AFM image of the process of rearranging single atoms at

room temperature. The image was acquired with a cantilever oscillation amplitude of 15.7 nm,

using a Si cantilever. (Reproduced from Ref. [96])

from one lattice site to the other they finally succeeded to write the letter “Sn”

with single atoms.

1.4

Summary

modern applications of atomic force microscopy. This versatile microscopy

technique can be categorized in two operational modes: static and dynamic

mode. The static mode allows the simultaneous measurement of normal and

lateral forces, thus yielding direct information about friction mechanisms of

nanoscale contacts. The main advantage of the dynamic mode is the possibil-

ity to control the tip-sample distance while avoiding the undesired since de-

structive jump-to-contact phenomenon. Two different excitation schemes for

dynamic force microscopy were introduced, where the amplitude-modulation

or tapping mode are in particular well-suited for high resolution imaging un-

der ambient or liquid conditions. The ultimate ’true’ atomic resolution, how-

ever, is limited to vacuum conditions using the frequency-modulation or non-

contact technique. Still, the impact of atomic force microscopy reaches far

beyond high resolution imaging of surface topography. The method of force

spectroscopy allows quantification of tip-sample forces, through the system-

atic acquisition of parameters like amplitude, phase and oscillation frequency

as a function of the relative tip-sample distance. Based on this approach not

only the bonding force of single interatomic chemical bonds can be measured,

but also the full 3-dimensional force field with atomic resolution can be de-

termined. Finally, atomic forces can not only be measured, but they can be

Bibliography 37

exerted with atomic precision opening the new and exciting field of nanoma-

nipulation.

The authors would like to thank all colleagues who contributed to this work

with their images and experimental results: Boris Anczykowski, Daniel Ebel-

ing, Jan-Erik Jan-Erik Schmutz, Domenique Weiner (University of Münster),

Wolf Allers, Shenja Langkat, Alexander Schwarz (University of Hamburg) and

Udo D. Schwarz (Yale University).

Bibliography

G. Couturier. Nonlinear dynamic behav- an insulator via ultrahigh vacuum dy-

ior of an oscillating tip-microlever system namic force microscopy. Probe Microscopy,

and contrast at the atomic scale. Phys. Rev. 1:3, 1997.

Lett., 82:3388–3391, 1999. 9 C. Barth and M. Reichling. Imaging

2 T. R. Albrecht, P. Grütter, D. Horne, and the atomic arrangement on the high-

D. Rugar. Frequency modulation de- temperature reconstructed α-Al2 O3 (0001)

tection using high-Q cantilevers for en- surface. Nature, 414:54–57, 2001.

hanced force microscope sensitivity. J. 10 B. Bhushan and O. Marti. Scanning probe

Appl. Phys., 69:668–673, 1991. microscopy - principle of operation, in-

3 S. Alexander, L. Hellemans, O. Marti, strumentation, and probes. In B. Bhushan,

J. Schneir, V. Elings, and P. K. Hansma. An editor, Nanotribology and Nanomechanics –

atomic-resolution atomic-force microscope An Introduction, pages 41–115. Springer-

implemented using an optical lever. J. Verlag Berlin Heidelberg, 2005.

Appl. Phys., 65:164–167, 1988. 11 G. Binnig, C. F. Quate, and Ch. Gerber.

Atomic force microscopy. Phys. Rev. Lett.,

4 W. Allers, S. Langkat, and R. Wiesendan-

56:930–933, 1986.

ger. Dynamic low-temperature scanning

force microscopy on nickel oxide (001). 12 G. Binnig, H. Rohrer, C. Gerber, and

Appl. Phys. A, 72:S27, 2001. E. Weibel. Surface studies by scaning

tunneling microscopy. Phys. Rev. Lett.,

5 W. Allers, A. Schwarz, U. D. Schwarz, and

49:57–61, 1982.

R. Wiesendanger. A scanning force micro-

scope with atomic resolution in ultrahigh 13 S.A. Burke, J.M. Mativetsky, R. Hoffmann,

vacuum and at low temperatures. Rev. Sci. and P. Grẗter. Nucleation and submono-

Instr., 69:221–225, 1998. layer growth of C6 0 onKBr. Phys. Rev.

Lett., 94:096102, 2005.

6 B. Anczykowski, D. Krüger, and H. Fuchs.

14 N. A. Burnham and R. J. Colton. Measur-

Cantilever dynamics in quasinoncontact

ing the nanomechanical properties and

force microscopy: Spectroscopic aspects.

surface forces of materials using an atomic

Phys. Rev. B, 53:15485–15488, 1996.

force microscopy. J. Vac. Sci. Technol. A,

7 Makoto Ashino, Alexander Schwarz, 7:2906, 1989.

Timo Behnke, and Roland Wiesendan- 15 J. P. Cleveland, B. Anczykowski, A. E.

ger. Atomic-resolution dynamic force Schmid, and V. B. Elings. Energy dis-

microscopy and spectroscopy of a single- sipation in tapping-mode atomic force

walled carbon nanotube: Characterization microscopy. Appl. Phys. Lett., 72:2613,

of interatomic van der waals forces. Physi- 1998.

cal Review Letters, 93:136101, 2004.

16 B. V. Derjaguin, V. M. Muller, and Y. P.

8 M. Bammerlin, R. Lüthi, E. Meyer, Toporov. Effect of contact deformations

A. Baratoff, J. Lü, M. Guggisberg, Ch. on the adhesion of particles. J. Colloid

Gerber, L. Howald, and H.-J. Güntherodt. Interface Sci., 53:314–326, 1975.

38 Bibliography

tion force and frequency shift in large- croscopy, 67:77–86, 1999.

amplitude dynamic force microscopy. 30 H. Hölscher. Quantitative measurement of

Appl. Phys. Lett., 75:433–435, 1999. tip-sample interactions in amplitude mod-

18 U. Dürig. Extracting interaction forces ulation atomic force microscopy. Appl.

and complementary observables in dy- Phys. Lett., 89:123109, 2006.

namic probe microscopy. Appl. Phys. Lett., 31 H. Hölscher, D. Ebeling, and U. D.

76:1203–1205, 2000. Schwarz. Theory of Q-controlled dy-

19 U. Dürig. Interaction sensing in dynamic namic force microscopy in air. J. Appl.

force microscopy. N. J. of Phys., 2:5.1–5.12, Phys., 99:084311, 2006.

2000. 32 H. Hölscher, B. Gotsmann, W. Allers, U. D.

Schwarz, H. Fuchs, and R. Wiesendanger.

20 R. Garcia, C. J. Gómez, N. F. Martinez,

Measurement of conservative and dissi-

S. Patil, C. Dietz, and R. Magerle. Identi-

pative tip-sample interaction forces with

fication of nanoscale dissipation processes

a dynamic force microscope using the fre-

by dynamic atomic force microscopy.

quency modulation technique. Phys. Rev.

Phys. Rev. Lett., 97:016103, 2006.

B, 64:075402, 2001.

21 R. Garcia and R. Pérez. Dynamic atomic 33 H. Hölscher, B. Gotsmann, W. Allers, U. D.

force microscopy methods. Surf. Sci. Rep., Schwarz, H. Fuchs, and R. Wiesendanger.

47:197–301, 2002. Comment on “damping mechanism in dy-

22 R. García and A. San Paulo. Dynam- namic force microscopy". Phys. Rev. Lett.,

ics of a vibrating tip near or in intermit- 88:019601, 2002.

tent contact with a surface. Phys. Rev. B, 34 H. Hölscher, B. Gotsmann, and

61:R13381–R13384, 2000. A. Schirmeisen. On dynamic force spec-

23 F.-J. Giessibl. Atomic resolution of the troscopy using the frequency modulation

silicon (111)-(7x7) surface by atomic force technique with constant excitation. Phys.

microscopy. Science, 267:68, 1995. Rev. B, 68:153401, 2003.

35 H. Hölscher, S. M. Langkat, A. Schwarz,

24 F.-J. Giessibl. A direct method to calculate

and R. Wiesendanger. Measurement

tip-sample forces from frequency shifts in

of three-dimensional force fields with

frequency-modulation atomic force mi-

atomic resolution using dynamic force

croscopy. Appl. Phys. Lett., 78:123–125,

spectroscopy. Appl. Phys. Lett, 81:4428–

2001.

4430, 2002.

25 F.-J. Giessibl. Advances in atomic force

36 H. Hölscher and A. Schirmeisen. Dynamic

microscopy. Rev. Mod. Phys., 75:949–983,

force microscopy and spectroscopy. In

2003.

P. W. Hawkes, editor, Advances in Imaging

26 F. J. Giessibl. Forces and frequency shifts and Electron Physics, pages 41–101, Lon-

in atomic-resolution dynamic-force mi- don, 2005. Academic Press Ltd.

croscopy. Phys. Rev. B, 56:16010–16015, 37 H. Hölscher, A. Schwarz, W. Allers, U. D.

Dec 1997. Schwarz, and R. Wiesendanger. Quan-

27 P. Gleyzes, P. K. Kuo, and A. C. Boccara. titative analysis of dynamic force spec-

Bistable behavior of a vibrating tip near a troscopy data on graphite(0001) in the

solid surface. Appl. Phys. Lett., 58:2989– contact and non-contact regime. Phys. Rev.

2991, 1991. B, 61:12678–12681, 2000.

28 B. Gotsmann, B. Ancykowski, C. Seidel, 38 H. Hölscher, U. D. Schwarz, and

and H. Fuchs. Determination of tip- R. Wiesendanger. Calculation of the fre-

sample interaction forces from measured quency shift in dynamic force microscopy.

dynamic force spectrosocpy curves. Appl. Appl. Surf. Sci., 140:344–351, 1999.

Surf. Sci., 140:314–319, 1999. 39 Frank C. Hoppenstaedt. Analysis and Sim-

ulation of Chaotic Systems. Springer-Verlag,

29 G. Haugstad and R. R. Jones. Mechanisms

New York, 1993.

of dynamic force microscopy on polyvinyl

alcohol: region-specific non-contact and 40 J. N. Israelachvili. Intermolecular and Sur-

face Forces. Academic Press, London, 1992.

Bibliography 39

bridge University Press, Cambridge, UK, 66:115409, 2002.

1985. 53 J. Legleiter and T. Kowalewski. Insight

42 K. L. Johnson, K. Kendall, and A. D. into fluid tapping-mode atomic force mi-

Roberts. Surface energy and contact of croscopy provided by numerical simula-

elastic solids. Proc. R. Soc. London Ser. A, tions. Appl. Phys. Lett., 87:163120, 2005.

324:301, 1971. 54 J. Legleiter, M. Park, B. Cusick, and

43 H. Kawakatsu, S. Kawai, D. Saya, M. Na- T. Kowalewski. Scanning probe accel-

gashio, D. Kobayashi, H. Toshiyoshi, eration microscopy (SPAM) in fluids:

and H. Fujita. Towards atomic force Mapping mechanical properties of sur-

microscopy up to 100 MHz. Review of faces at the nanoscale. Proc. Natl. Acad.

Scientific Instruments, 73(6):2317–2320, Sci. U.S.A., 103:4813–4818, 2006.

2002. 55 R. Linnemann, T. Gotszalk, I. W.

44 D. Klinov and S. Maganov. True molec- Rangelow, P. Dumania, and E. Oester-

ular resolution in tapping-mode atomic schulze. Atomic force microscopy and

force microscopy with high-resolution lateral force microscopy using piezore-

probes. Appl. Phys. Lett., 84:2697–2698, sistive cantilevers. J. Vac. Sci. Technol. B,

2004. 14(2):856–860, 1996.

45 A. Kühle, A. Sorensen, and J. Bohr. Role 56 Ch. Loppacher, M. Bammerlin, F. Battis-

of attractive forces in tapping tip force ton, M. Guggisberg, D. Müller, R. Lüthi

microscopy. J. Appl. Phys., 81:6562–6569, H. R. Hidber, E. Meyer, and H.-J. Gün-

1998. therodt. Fast digital electronics for ap-

plication in dynamic force microscopy

46 A. Kühle, A. H. Sørensen, J. B. Zandber-

using high-q cantilevers. Appl. Phys. A.,

gen, and J. Bohr. Contrast artifacts in

66:S215–S218, 1998.

tapping tip atomic force microscopy. Appl.

Phys. A, 66:S329–S332, 1998. 57 Ch. Loppacher, M. Guggisberg, O. Pfeif-

fer, E. Meyer, M. Bammerlin, R. Luthi,

47 T. Kunstmann, A. Schlarb, M. Fendrich,

R. Schlittler, J. K. Gimzewski, H. Tang,

Th. Wagner, R. Möller, and R. Hoffmann.

and C. Joachim. Direct determination

Dynamic force microscopy study of

of the energy required to operate a sin-

3,4,9,10-perylenetetracarboxylic dianhy-

gle molecule switch. Phys. Rev. Lett.,

dride on kbr(001). Phys. Rev. B, 71:121403,

90:066107, 2003.

2005.

58 Ch. Loppacher, U. Zerweck, L.M. Eng,

48 L. D. Landau and E. M. Lifschitz. Lehrbuch

S. Gemming, G. Seifert, C. Olbrich,

der Theoretischen Physik: Mechanik.

K. Morawetz, and M. Schreiber. Ad-

Akademie-Verlag, Berlin, 1990.

sorption of PTCDA on a partially KBr cov-

49 L. D. Landau and E. M. Lifschitz. Lehrbuch ered Ag(111) substrate. Nanotechnology,

der theoretischen Physik VII: Elastizitätstheo- 17:1568, 2006.

rie. Akademie-Verlag, Berlin, 1991. 59 R. Lüthi, E. Meyer, H. Haefke, L. Howald,

50 M. A. Lantz, H. Hug, R. Hoffmann, P. J. A. W. Gutmannsbauer, M. Guggisberg,

van Schendel, P. Kappenberger, S. Martin, M. Bammerlin, and H.-J. Güntherodt.

A. Baratoff, and H.-J. Güntherodt. Quanti- Nanotribology: an UHV-SFM study on

tative measurement of short-range chemi- thin films of C6 0 and AgBr. Surf. Sci.,

cal bonding forces. Science, 291:2580–2583, 338:247–260, 1995.

2001. 60 S. Maganov. Visualization of polymer

51 M. Lee and W. Jhe. General solution of structures with atomic force microscopy.

amplitude-modulation atomic force mi- In H. Fuchs, M. Hosaka, and B. Bhushan,

croscopy. Phys. Rev. Lett., 97:036104, 2006. editors, Applied Scanning Probe Methods,

52 S. I. Lee, S. W. Howell, A. Raman, and pages 207–250. Springer-Verlag, 2004.

R. Reifenberger. Nonlinear dynamics of 61 Y. Martin, C. C. Williams, and H. K. Wick-

microcantilevers in tapping mode atomic ramasinghe. Atomic force microscopeŰ-

force microscopy: a comparison between force mapping and profiling on a sub

40 Bibliography

(1987). and S. Morita. Mechanical vertical ma-

62 C. M. Mate, G. M. Mcclelland, R. Erlands- nipulation of selected single atoms by

son, and S. Chiang. Atomic-scale friction soft nanoindentation using near contact

of a tungsten tip on a graphite surface. atomic force microscopy. Phys. Rev. Lett.,

Phys. Rev. Lett., 59:1942 – 1945, 1987. 90:176102, 2003.

ment, and C. Abell. Chiral discrimination resolution imaging of antibodies by

by chemical force microscopy. Nature, tapping-mode atomic force microscopy:

391:566–568, 1998. Attractive and repulsive tip-sample inter-

action regimes. Biophys. J., 78:1599–1605,

64 E. Meyer, H.-J. Hug, and R. Bennewitz.

2000.

Scanning Probe Microscopy – The Lab on a

Tip. Springer-Verlag, 2004. 75 C. A. J. Putman, K. O. Vanderwerf, B. G.

Degrooth, N. F. Vanhulst, and J. Greve.

65 G. Meyer and N. M. Amer. Novel opti-

Tapping mode atomic force microscopy

cal approach to atomic force microscopy. in liquid. Appl. Phys. Lett., 64:2454–2456,

Appl. Phys. Lett., 53:1045–1047, 1988. 1994.

66 C. Möller, M. Allen, V. Elings, A. Engel,

76 T. R. Rodríguez and R. García. Tip motion

and D. J. Müller. Tapping-mode atomic in amplitude modulation (tapping-mode)

force microscopy produces faithful high- atomic-force microscopy: Comparison be-

resolution images of protein surfaces. tween continuous and point-mass models.

Biophys. J., 77:1150–1158, 1999. Appl. Phys. Lett, 80:1646–1648, 2002.

67 S. Morita and Y. Sugawara. In S. Morita,

77 D. Rugar, H. J. Mamin, and P. Guethner.

R. Wiesendanger, and E. Meyer, edi- Improved fiber-optic interferometer for

tors, Noncontact Atomic Force Microscopy, atomic force microscopy. Applied Physics

pages 47–77, Heidelberg, Germany, 2002. Letters, 55:2588–2590, 1989.

Springer-Verlag.

78 J. E. Sader. Parallel beam approximation

68 S. Morita, R. Wiesendanger, and E. Meyer,

for V-shaped atomic force microscope can-

editors. Noncontact Atomic Force Mi- tilevers. Review of Scientific Instruments,

croscopy. Springer-Verlag, Berlin, 2002. 66:4583–4587, 1995.

69 A. Moser, H.-J. Hug, T. Jung, U. D.

79 J. E. Sader and S. P. Jarvis. Accurate for-

Schwarz, and H.-J. Guntherodt. A minia- mulas for interaction force and energy in

ture fibre optic force microscope scan frequency modulation force spectroscopy.

head. Measurement Science and Technology, Appl. Phys. Lett., 84:1801–1803, 2004.

4:769–775, 1993.

80 J. E. Sader, T. Uchihashi, A. Farrell

70 J. M. Neumeister and W. A. Ducker. Lat-

M. J. Higgins, Y. Nakayama, and S.P.

eral, normal, and longitudinal spring Jarvis. Quantitative force measurements

constants of atomic force microscopy can- using frequency modulation atomic force

tilevers. Rev. of Sci. Instrum., 65:2527–2531, microscopy - theoretical foundations.

1994. Nanotechnology, 16:S94–S101, 2005.

71 L. Nony, R. Boisgard, and J.-P. Aimé.

81 O. Sahin, C. F. Quate, O. Solgaard, and

Stability criterions of an oscillating tip- A. Atalar. Resonant harmonic response

cantilever system in dynamic force mi- in tapping-mode atomic force microscopy.

croscopy. Eur. Phys. J. B, 24:221–229, 2001. Phys. Rev. B, 69:165416, 2004.

72 R. M. Overney, E. Meyer, J. Frommer,

82 A. San Paulo and R. García. Unifying

D. Brodbeck, R. Lüthi, L. Howald, H.- theory of tapping-mode atomic-force mi-

J. Güntherodt, M. Fujihira, H. Takano, croscopy. Phys. Rev. B, 66:041406, 2002.

and Y. Gotoh. Friction measurements on

phase-separated thin films with a mod- 83 D. Sarid. Scanning Force Microscopy –

ified atomic force microscope. Nature, With Applications to Electric, Magnetic, and

359:133–135, 1992. Atomic Forces. Oxford University Press,

1994.

Bibliography 41

84 Naruo Sasaki and Masaru Tsukada. The- 94 B. Such, D. Weiner, A. Schirmeisen, and

ory for the effect of the tip-surface inter- H. Fuchs. Influence of the local adsorption

action potential on atomic resolution in environment on the intra-molecular con-

forced vibration system of noncontact trast of organic molecules in non-contact

afm. Appl. Surf. Sci., 140:339–343, 1999. atomic force microscopy. Appl. Phys. Lett,

85 A. Schirmeisen, D. Weiner, and H. Fuchs. 89:093104, 2006.

Single-atom contact mechanics: From 95 Y. Sugawara, M. Otha, H. Ueyama, and

atomic scale energy barrier to mechani- S. Morita. Defect motion on an InP(110)

cal relaxation hysteresis. Phys. Rev. Lett., surface observed with noncontact atomic

97:136101, 2006. force microscopy. Science, 270:1646, 1995.

86 C. Schönenberger and S. F. Alvarado. A 96 Y. Sugimoto, M. Abe, S. Hirayama, N. Oy-

differential interferometer for force mi- abu, O. Custance, and S. Morita. Atom

croscopy. Rev. Sci. Instrum., 60:3131–3134, inlays performed at room temperature

1989. using atomic force microscopy. Nature

87 U. D. Schwarz. A generalized analytical Materials, 4:156–159, 2005.

model for the elastic deformation of an 97 J. Tamayo and R. García. Deformation,

adhesive contact between a sphere and a contact time, and phase contrast in tap-

flat surface. J. Coll. Interface Sci., 261:99– ping mode scanning force microscopy.

106, 2003. Langmuir, 12:4430–4435, 1996.

88 U. Stahl, C. W. Yuan, A. L. de Lozanne, 98 J. Tamayo and R. García. Relationship

and M. Tortonese. Atomic force micro- between phase shift and energy dissipa-

scope using piezoresistive cantilevers and tion in tapping-mode scanning force mi-

combined with a scanning electron mi- croscopy. Appl. Phys. Lett., 73:2926–2928,

croscope. Applied Physics Letters, 65:2878– 1998.

2880, 1994. 99 M. Tortonese, R. C. Barrett, and C. F.

89 M. Stark, R. W. Stark, W. M. Heckl, and Quate. Atomic resolution with an atomic

R. Guckenberger. Inverting dynamic force force microscope using piezoresistive de-

microscopy: From signals to time-resolved tection. Applied Physics Letters, 62:834–836,

interaction forces. PNAS, 99:8473–8478, 1993.

2002. 100 H. Ueyama, Y. Sugawara, and S. Morita.

90 R. W. Stark and W. Heckl. Fourier trans- Stable operation mode for dynamic non-

formed atomic force microscopy: tapping contact atomic force microscopy. Appl.

mode atomic force microscopy beyond Phys. A, 66:S295 – S297, 1998.

the hookian approximation. Surf. Sci., 101 L. Wang. Analytical descriptions of the

457:219–228, 2000. tapping-mode atomic force microscopy

91 R. W. Stark, G. Schitter, M. Stark, R. Guck- responce. Appl. Phys. Lett., 73:3781–3783,

enheimer, and A. Stemmer. State-space 1998.

model of freely vibrating and surface- 102 C. W. Yuan, E. Batalla, M. Zacher, A. L.

coupled cantilever dynamics in atomic de Lozanne, M. D. Kirk, and M. Tortonese.

force microscopy. Phys. Rev. B, 69:085412, Low temperature magnetic force micro-

2004. scope utilizing a piezoresistive cantilever.

92 R. W. Stark, G. Schitter, and A. Stemmer. Applied Physics Letters, 65:1308–1310, 1994.

Tuning the interaction forces in tapping 103 Q. D. Zhong, D. Inniss, K. Kjoller, and

mode atomic force microscopy. Phys. Rev. V. B. Elings. Fractured polymer/silica

B, 68:085401, 2003. fiber surface studied by tapping mode

93 Th. Stifter, O. Marti, and B. Bhushan. The- atomic force microscopy. Surf. Sci. Lett.,

orectical investigation of the distance de- 290:L688–L692, 1993.

pendence of capillary and van der waals 104 L. Zitzler, S. Herminghaus, and F. Mugele.

forces in scanning force microscopy. Phys. Capillary forces in tapping mode atomic

Rev. B, 62:13667–13673, 2000. force microscopy. Phys. Rev. B, 66:155436,

2002.

43

List of Contributors

Westfälische Wilhelms-Universität Mün- Westfälische Wilhelms-Universität Mün-

ster ster

Center for NanoTechnology (CeNTech) Physikalisches Institut

Heisenbergstr. 11 Wilhelm-Klemm-Str. 10

48149 Münster 48149 Münster

Germany Germany

Hendrik.Hoelscher@uni-muenster.de Schirmeisen@uni-muenster.de

• Figure 1.20 is courtesy of Shenja Langkat and Wolf Allers, Hamburg Uni-

versity

Copyright © 2006 Wiley-VCH, Weinheim (Germany)

ISBN: 3-527-XXXXX-X

6.5 Weiterführende Literatur

Scanning Probe Microscopy and Spectroscopy: Methods and Applications, Roland Wiesendanger,

Cambridge University Press, 1994.

Scanning Force Microscopy: With Applications to Electric, Magnetic, and Atomic Forces, Dror Sarid,

Oxford University Press, 1991.