The JRP aims to develop metrological traceability for the measurement of the mechanical properties (adhesion, stiffness and elasticity) of nano-objects, such as nanoparticles, nanowires and nanoscale structures and composite materials. Forces from 10 mN down to 10 pN will be applied to objects with sizes down to 50 nm.
Background
Nano-objects are used in industry to improve the properties and functionality of everyday products like glass, steel, cement, coatings and easy-to-clean ceramics and glass. The dimensions of nano-objects can be measured using Atomic Force Microscopy (AFM). The measurement principle of the AFM, using thin silicon beams with integrated probing tips at the free end and detecting the deflection of the cantilever during the measurement, allows forces to be measured. This opens up a completely new field of force spectroscopy applications for AFMs. One of these application fields is the measurement of the mechanical properties of nano-objects. Several approaches to measure the elastic modulus of nano-objects have been published in the past 10 years. The scatter and deviation of the results amount to more than 100 % meaning that at the moment AFM-methods are used to map elasticity rather than measure it. A further problem arises when measuring the mechanical properties of nano-objects. No reliable methods exist which can be used to tether nano-particles or nano-wires to a stiff substrate. Furthermore the interaction forces between tip and sample disturb the measurement of nanomechanical properties. No methods exist to date to minimize these influences on the measurement results. There is also recent evidence to suggest that a change of the elastic modulus occurs with changes with the size of nano-objects. Due to the lack of traceability this cannot be systematically investigated.
Need for the project
The mechanical properties of nano-objects are increasingly attractive for industry, since nano-objects can be used to improve the properties, functionality and lifetime of everyday products. A prerequisite for reliable mechanical property measurements is the fixation of nano-objects during contact measurements. Atomic force microscopy (AFM) with its capability of measuring the topography of nano-objects can also be used to measure their mechanical properties. Several methods using AFMs have been developed but suffer from being not verified and traceable to the SI units. Furthermore the influence of tip–sample interaction, sample–base interaction, surface impurities and environmental influences like humidity also lead to systematic deviations. However, since at the nanoscale not all effects are observable it is mandatory to support experimental results by modelling and simulation. Available methods are Finite Element Modelling (FEM) and Molecular Dynamic Calculations (MDC). The former needs to be extended towards smaller dimensions (1 µm) and the latter towards larger dimensions (100 nm). Comparative measurements between the well known instrumented indentation testers and AFMs on the developed test samples need to be carried out, in order to quantify measurement uncertainties of the developed AFM methods of bending and contact-resonance frequency-modulation (CR-FM).
Scientific and technical objectives
The project has the following scientific and technical objectives:
• Development of traceable stiffness calibration of AFM cantilevers in the force range of 1 µN to 10 pN
• Provision of a method for easy-to-use in-situ calibration of AFM cantilever stiffness and one for instrumented indenters
• Provision of three test structures with arrays of structures for a qualification of instrumented indentation technique (IIT) and AFM
• Provision of Finite Element and Molecular Dynamic simulation tools to model the deformations of nano-objects from molecular level to object sizes of 100 nm to support experimental data
• Performing an interlaboratory comparison on a set of test samples to reveal their mechanical properties as a first step into the creation of new types of reference materials
• Provision of a guideline for the measurement of the geometrical parameters, e.g. length, diameter, width and height of nanostructures and nano-objects and to establish correction factors considering tip–sample effects
• Provision of a guideline for the preparation of fixed nano-objects on hard substrates, like nanoparticles, and nanowires
• Provision of a guide line for the measurement of the mechanical properties of small nanosized objects by AFM and IIT
• Provision of uncertainty budgets related to the measurement of mechanical properties, such as elastic modulus.
Results and potential impact
The primary task of investigating nano-objects and biomaterial is to unravel new material phenomena to exploit them for new materials functions and to optimize them for improved or new products. The mechanical properties of nano-objects are key elements for their application in improved or new products in a very large variety. The Young’s modulus for example of a single graphite sheet is about 1.3 TPa, those of single wall carbon nano tubes (SW-CNTs) is about 1 TPa still making CNT one of the strongest materials. CNTs are already used in many applications to improve classical products, e.g. by adding to concrete to improve its properties, or a single CNT as a high-resolution, low wear tip in SFM.
Improved mechanical functions of nano-objects are not restricted to one class of products, e.g. IC only. As stated in the Roadmap of the EuMat “New materials technological solutions serve also as a catalyst for other industries being an enabling technology for new innovations in process, machine building, energy, transportation and aeronautics industries. Advanced life cycle cost based solutions strengthen the technology and competitiveness of European after-market service providers. Nano-(bio-)materials are related for example to lighter materials with higher strength to reduce energy consumption or for safety protection, as thin hard films to improve scratch resistance of surfaces, as thin membranes with nanoholes to be used for purification of water and other liquids or energy production, or used as small actuators and sensors in nanorobots, piezoelectric and semiconducting nanowires or arrays (e.g. ZnO) which are capable of mechanical to electrical transduction to generate electrical power being deformed elastically by direct contact or by vibration useful for energy harvesting devices.
"Moreover the enhancement of mechanical and thermal properties of polymers by the inclusion of inorganic moieties, especially in the form of nanocomposites, offers the possibility to substitute metals or traditional composites in the automotive industry with innovative hybrid materials”. A better knowledge of the measurement of mechanical properties of nano-objects and biomaterials combined with reliable and comparable results would improve the transfer to further industrial application.
The impact of the project is related to the provision of methodologies, calibration and test tools, and improved analytical tools for modelling and simulation, and with 
best practice guides for improved measurement of size, shape and mechanical properties of nanoscale objects. This will enable the research at institutes and industry level making better selection of the appropriate methods, and provide faster validation of new materials.
Development of traceable stiffness calibration of AFM cantilevers in the force range of 1 µN to 10 pN
Together with the collaborator NanoWorld Services GmbH PTB improved the Thermal Tune method for the calibration of the bending stiffness of AFM cantilevers with stiffnesses ranging from 0.1 N/m to 5 N/m. NanoWorld uses calibrated reference cantilevers to calibrate their Thermal Tune method. PTB calibrated the stiffness of two reference cantilevers for NanoWorld. The stiffness uncertainty of these two cantilevers amounted to 4 %. NanoWorld states an uncertainty of their improved Thermal Tune method of 10 %. In order to confirm the quality of the new method comparison measurements between PTB and NanoWorld of three different types of cantilevers (PPP NCST-3, PPP FM-3 and PPP Cont-3) were carried out. Three cantilevers of each type were measured. The deviations of NanoWorlds values from the PTB values for the three types of cantilevers were 7.6 %, 5.7 % and 7.3 % with an average value of (6.9 ± 2.9) %. These results confirm NanoWorlds stated uncertainty of 10 % for the normal stiffness of AFM cantilevers.
Stiffer cantilevers (> 5 N/m) can be calibrated by PTB applying its compensation balance calibration setup with a minimum uncertainty of 4 %. The smallest forces applied during the calibration are determined by the minimum deflection needed for calibration (30 nm). This leads for stiffnesses of 5 N/m to forces of 150 nN and for stiffnesses of 50 N/m to forces of 1500 nN. If these calibrations are carried out applying the load to the AFM tip, then tip damages of tips with nm radius are very probable leading to systematic deviations of the measured stiffness of some ten percent. To avoid this error it is planned to extend PTB’s calibration service to the backside of cantilevers. Thus the user has to correct the cantilever stiffness measured at the end of the cantilever to the position of the AFM tip apex.
AFM Cantilever stiffnesses kCant below 0.1 N/m can be calibrated in future using a new calibration setup of the PTB based on calibrated MEMS reference springs (kMEMS) [Gao10]. One constraint of this method is that reference springs are needed which lie within the range of 0.1 < kCant/kMEMS < 10. This setup has been tested in the stiffness range from 0.1 N/m to 40 N/m and provides an uncertainty in stiffness of 7 %. For cantilevers with smaller stiffness new MEMS reference springs with a stiffness of 0.5 N/m were produced which are currently under test. With these reference springs it is expected that cantilever stiffnesses down to 0.05 N/m can be calibrated.
Provision of a method for easy-to-use in-situ calibration of AFM cantilever stiffness and one for instrumented indenters
PTB offers one common solution for both, in-situ stiffness calibration of AFM cantilevers and instrumented indenters. The method is based on the well-known cantilever on reference spring method [5]. PTB developed together with the Technical University of Braunschweig and the CiS GmbH meander springs with a stiffness of 15 N/m. The linearity deviations up to 1 mN are smaller than 0.2 % and more important, the influence of the loading position on the stiffness is strongly reduced (< 1 %). The reference springs are available from CiS GmbH in Erfurt (Germany) and PTB offers the calibration of their stiffness with an uncertainty of 4 %.
A second approach developed by PTB, not only allows to calibrate stiffness, but at the same time force and deflection. The idea originally designed for instrumented indenters in the force range up to 500 mN is here realized for forces up to 12 µN and displacements up to 3 µm. These new MEMS reference springs show no dependence of their stiffness on loading position on the MEMS loading area (50 µm x 50 µm) and moreover the temperature dependence of stiffness, bending force and bending deflection is for ambient conditions below 1 %. REG TUCh is aiming at offering these MEMS reference springs with stop limit to customers in the near future.
Provision of three test structures with arrays of structures for a qualification of instrumented indentation technique (IIT) and AFM
The MechProNO partners provide arrays of pillars, cantilevers and nano-particles for a qualification of AFM and instrumented indenters. Two different pillar materials were fabricated, silicon and photo-resist. IIT measurements on high aspect ratio silicon pillars revealed that the equivalent stiffness of the pillars has to be taken into account [2]. It is proportional to the square of the pillar radius and inversely proportional to the pillar height. Both, diameter and height were measured using the traceable dimensional AFM of NPL. If the equivalent pillar stiffness is taken into account, then comparable indentation moduli compared to the bulk material resulted. Ignoring this effect leads to decreasing indentation moduli down to 35 % of the bulk value.
The manufacturing of cantilevers proved to be complicated due to the fact that the cantilever stiffness depends on the cantilever thickness to the third power. FIB production of cantilevers out of the bulk material [3] lead to thickness variations of the cantilevers of up to 50 %. Alternatively cantilevers manufactured by FIB processing of membranes and by wet chemically etching were successfully tested. The thickness of the membranes could be measured by x-ray diffraction. AFM topography measurements on the cantilevers in contact mode allows to check the force calibration of the AFM.
Samples with a defined uniform distribution of nano-particles could be realized by REG TUD. For gold nano-rods (25 nm x 77 nm) the indentation process of an AFM tip could be simulated by REG UH using the molecular dynamics technique. AFM indentation measurements are underway to test the developed model.
A new technique using standard TEM lamellae for bridge/beam fabrication by FIB was tested by BAM. At least for one material, amorphous SiO2, appropriate samples enabling IIT-measurements on increasingly smaller objects could be realized. Thus, 2 µm thick and 30 nm thick silica layers on Si were measured by IIT. Furthermore, pillars of 2 µm height with diameters ranging from 0.5 µm to 2 µm were prepared from the same material.
A specimen containing 20 nm spherical silica nano-particles in Epoxy has been prepared and is now available for measurements.
For AFM, beams of different size from 0.3 - 5 µm were cut by FIB from the free standing 30 nm thick silica membrane.
Provision of Finite Element and Molecular Dynamic simulation tools to model the deformations of nano-objects from molecular level to object sizes of 100 nm to support experimental data
Atomistic model systems at a scale of 1:2 of gold nanorods and a diamond AFM tip apex (rtip = 5/10 nm) have been built and molecular dynamics (MD) simulations of AFM nanoindentation experiments were carried out [4]. The aim of these simulations was to test and validate the simulation technique and obtain some results on the mechanical properties of gold nanorods. The actual samples used in experiment consist of commercial gold nanorods (Nanopartz Bare Gold Nanorodz items 30-25-600 and 30-25-700), deposited on silicon or silica substrates. Two different crystallographic structures for the nanorods were studied: two single crystal, one grown along the <100> direction and the other along the <110> direction, with an octagonal cross-section and one penta-twinned crystal grown along <100> direction, corresponding to the structure of the sample as specified by the manufacturer.
The system is highly idealized. It assumes clean surfaces of tip, nanorod, and substrate, no humidity and the indentation rate is assumed to be 109 times faster than in real experiments. Moreover the experimental conditions are far from UHV. Real surfaces are corrugated, and adsorbants may be present. A thin water layer on the sample leads to the formation of a meniscus as the AFM tip approaches the surface.
Force-distance curves were achieved showing
• after short “elastic” regime, dislocations nucleate in the nanorods and move in {111} glide planes towards the free surfaces
• when dislocations annihilate at surfaces, stress is released and the force drops
• in the single crystal nanorods, the orientation affects the deformation mechanism
• in the penta-twinned nanorod, the dislocations are trapped at the twinning planes
• upon retraction of the tip, vacancies remain under the area of indentation in the single crystal rods. In the penta-twinned rod, the dislocations remain trapped at the twinning planes
• in all systems, the deformation is irreversible
Performing an interlaboratory comparison on a set of test samples to reveal their mechanical properties as a first step into the creation of new types of reference materials
PTB has investigated the mechanical properties of silicon nano-pillars with different diameters using the nanoindentation technique. The height and diameter of these chemically etched nanopillars have been measured using atomic force microscopy (AFM). Experimental results for two pillars, one with 386 nm diameter (aspect ratio of 1.3) and one with 2 µm diameter, show a decrease of the indentation modulus by 50 % for the small diameter pillar and a reduction of 20 % for the bigger diameter. Taking into account the compressibility of the pillars, the measured indentation modulus can be corrected and modulus values comparable to the bulk material are achieved (see [4]). Thus no size effect was observed for Silicon nanopillars.
PTB calibrated a set of cantilevers which were calibrated by NanoWorld using the improved thermal tune method. PTB used the MEMS reference spring method and the direct calibration of stiffness using a compensation balance. VTT has performed cantilever spring constant measurements with the Thermal Noise method, the Sader method and direct force measurements with a balance. Result of the measurements was that AFM tip damage was observed due to too high contact forces during calibration. Tip damage usually leads to plateau like tips. Probing on different positions of these plateaus leads to varying stiffnesses of up to 20 %.
Provision of a guideline for the measurement of the geometrical parameters, e.g. length, diameter, width and height of nanostructures and nano-objects and to establish correction factors considering tip–sample effects
NPL has set-up an AFM for studying the tip-sample interaction. The system features upgraded hardware and software that enable scans to be performed with a noise level lower than 0.3 nm. NPL confirmed the instrument works correctly by measuring calibrated height standards. A 108 nm step height was measured to have a height of 107.7 nm ± 0.3 nm. It was also verified that the system works correctly both in contact and non-contact mode, with no perceived variation of the results depending on the scanning parameters (e.g. non-contact mode oscillation amplitude).
From TUD, NPL received samples made of several nanoparticles dispersed on Si or SiO2 substrates. The results show the average diameter of the several nanoparticles is in agreement with the nominal value, but there is quite a large standard deviation from the average value (up to 15 %). These results are in agreement with the ones recently reported by TUD. PTB also provided NPL with samples of silicon or photoresist pillars on silicon substrates. NPL’s measurements showed that the average height of the pillars does not correspond to the nominal values; there is a small deviation (< 2 %). We attribute the discrepancy between the nominal height and the measured average height to the fabrication process.
NPL also accomplished the implementation of the blind tip reconstruction routine for reconstructing the shape of the tip from AFM images. The software features the possibility of manually filtering out part of the noise of the measurement, or automatically identifying the optimum noise filtering parameter. The software and a short manual are available to the partners and it is being tested on results from several samples and AFMs.
NPL has conducted experiments on the effect of tip wear and tip material on AFM measurements of silver (Ag) nanoparticles (NPs) on Si or SiO2 substrates. Starting with the effect of tip wear, they used the AFM in contact mode (CM) and non-contact mode (NCM) with silicon tips. For both cases they first measured a group of Ag nanoparticles dispersed on a SiO2 substrate, then they eroded the tip by pressing it against an area free of NPs while scanning, and finally they took AFM measurements of the same nanoparticles. A direct comparison between the NP’s height measured before and after eroding the tip was made. The results show an increase in height of an average of 4 nm in the case of contactmode (CM) and a reduction of the measured height of 1 nm for NCM. We attribute these effects to the different interaction between tip/sample and tip/NP depending on the tip radius. This causes for example a variation of the AFM working distance from the substrate/NP due to an increase of the adhesion forces for larger radii as well as a larger compression of the NP with a smaller radius.
NPL also made AFM measurements of Ag NPs taken with two different kinds of tips (specifically silicon or diamond tips) working in NCM with the same free space oscillation and contact set-point. In this case the height of the NPs when dispersed on either a silicon substrate or a silicon dioxide substrate was compared. The aim is to identify if the different substrate or tip material induces a different measurement of the height of the NPs. The results showed a reduction of ~1 nm in the average of the measured height of the NP, when using a diamond tip instead of a silicon tip.
Provision of a guideline for the preparation of fixed nano-objects on hard substrates, like nanoparticles, and nanowires
REG TUD has delivered the following sets of nanoparticle samples prepared on Si or SiO2 substrates to the partners:
• Gold nanoparticles (60 nm, spherical)
• Gold nanorods 650 and 700 (25 nm x 66 nm and 25 nm x 77 nm respectively)
• Silica nanoparticles (100 nm, respectively 304 nm, spherical)
• Silver nanoparticles (50 nm, spherical)
• Titanium dioxide (<150 nm)
Guidelines for optimised preparation methods of fixed nano-objects on substrates as drying, rinsing, dip coating and electrostatic precipitation were developed and their fitness was evaluated. To deposit nano-objects in a defined distance to each other a model for calculating the particle area concentration depending on particle sizes has been developed. Furthermore suitable procedures for substrate cleaning as wet chemical cleaning and dry cleaning methods were investigated and all of the selected materials were measured concerning their particle size or surface charge with laboratory instruments.
Provision of a guide line for the measurement of the mechanical properties of small nanosized objects by AFM and IIT
VTT has manufactured new samples of gold nanorods with nominal dimensions of 25 nm in diameter and 73 nm in length. The nanorods are better distinguishable with AFM than the earlier ones and thus are very suitable for force-distance measurements.
VTT has investigated the noise and temperature drift properties of its humidity-controlled AFM. Tests of adhesion force measurements with a commercial AFM have been carried out. Moreover a lot of force-distance measurements to understand more about the force measurements with AFM were made under ambient laboratory conditions.
BAM finished the work on the revision of ISO 14577 parts 1 to 3 which describes the measurement of mechanical properties by instrumented indentation. This guide line is applicable for measuring the mechanical properties of small microsized objects by IIT. Nanosized objects cannot be localized by instrumented indentation instruments and thus measurements are not possible.
Literature
[1] S. Gao, Z. Zhang, Y. Wu, und K. Herrmann, „Towards quantitative determination of the spring constant of a scanning force microscope cantilever with a microelectromechanical nano-force actuator“, Meas. Sci. Technol., Bd. 21, Nr. 1, S. 015103, Jan. 2010.
[2] Z. Li, S. Gao, F. Pohlenz, U. Brand, L. Koenders, und E. Peiner, „Determination of the mechanical properties of nano-pillars using the nanoindentation technique“, Nanotechnol. Precis. Eng., Bd. 3, S. 182–188.
[3] N. Wollschläger, W. Österle, I. Häusler, und M. Stewart, „Ga+ implantation in a PZT film during focused ion beam micro-machining“, Phys. Status Solidi C, Bd. 12, Nr. 3, S. 314–317, März 2015.
[4] Reischl B, Kuronen A and Nordlund K 2014 Nanoindentation of gold nanorods with an atomic force microscope. Mater. Res. Express 1, 4, 045042
[5] Tortonese M, Kirk M 1997 Characterization of application specific probes for SPMs Proc. SPIE 3009 53–60
For more information and if you are interested in some of the nano-objects developed in the course of this project, please contact: 
Uwe Brand