You are on page 1of 4

Measurement: Sensors 18 (2021) 100109

Contents lists available at ScienceDirect

Measurement: Sensors
journal homepage: www.sciencedirect.com/journal/measurement-sensors

Metrological atomic force microscope for calibrating nano-scale step height standards

A R T I C L E I N F O A B S T R A C T

Keywords In this paper, we described a metrological AFM in order to calibrate step height standards used as transfer ar­
He-Ne laser Displacement interferometer tefacts for commercial AFMs. With X, Y and Z axes laser displacement interferometers installed, related cali­
Atomic force microscope (AFM) brated results are directly traceable to the SI of metre definition. The step height measurement accuracy was
Nano-scale step height standards
verified using several step heights up to 2000 nm. Corresponding measurement uncertainty in step height
Measurement uncertainty analysis
standard calibration was estimated in line with the Guide to the Expression of Uncertainty in Measurement
(GUM). The expanded measurement uncertainty (k = 2) was better than 2.5 nm over a full range of up to 2000
nm. This suggests that the developed metrology system is capable of conducting nano-scale metrological stan­
dard calibration which is applicable to disseminating accuracy and traceability to commercial AFMs and optical
microscopes in application of surface topography measurement.

1. Introduction topography scales are typically calculated by the voltages applied on the
piezo-electric transducer (PZT) stage or PZT tube to result in a relative
Atomic Force Microscopy (AFM) [1] has already become a popular linear scanning motion of a sharp AFM tip and the test sample surface.
measurement means to characterize the properties of surfaces down to Basically, the feedback sensors, e.g., capacitive sensor, linear voltage
the nanometre scale and hence are adopted to be of importance tool in differential transducers (LVDT), strain gauge and optical sensors, are
the building up nanometrology infrastructure. There are a number of used in the close-loop control for determining those scales as well as
standard artefacts available to do calibration of AFM scanning image accuracy of the stage movement in the order of hundred micrometers for
scales. In order to properly calibrate X-axis and Y-axis scales in AFMs, stage and tens of micrometers for the tube. The calibration of those
either lateral 1-dimensinal or 2-dimensional regular structures are scales of PZT could be implemented by the use of pre-calibrated refer­
usually employed. To conduct calibration of the Z axis scale, a series of ence standard samples/artefacts. However, due to the inherent natures
known step heights are usually being used. To do accurate calibration on of PZT, such as nonlinearity, hysteresis, drifting, creep and aging effects,
those artefacts, a metrological-AFM with laser interferometers installed the resulting image is always distorted. Obviously, from metrological
is necessary to provide corresponding reference values on those step point of view, it is noted that there is no good enough accuracy obtained
height standards in the area of nanometrology so as to enable the by only doing calibration of the scales of the PZT with feedback sensors,
measurement to have a direct measurement traceability to SI unit of the especially in a large scanning range up to millimetres. Therefore, it is
metre, which is defined as the length of path travelled by light in vacuum necessary to take laser interferometer as a displacement sensor to be
during a time interval of 1/299 792 458 of a second. integrated into each scanning axis of the motion stage to provide a direct
AFMs for nanometrology have been researched widely by national traceability of the measurement/calibration in line with the national
metrology institutes to build up metrology infrastructure to enable ac­ standard of length ─ definition of the metre.
curate step height standards capabilities [2–6]. National Metrology
Centre (NMC), A*STAR, Singapore, has also developed a
metrological-AFM [7] primarily to conduct measurement and calibra­ 2.1. Description of AFM probe and 3D nano-scale moving platform
tion of those step height standards to transfer accurate traceable cali­
bration to other normal AFMs. In this paper, the system’s configuration An AFM tiny tip in a nominal radius is less than 10 nm attached onto
and calibration implementation on nanoscale step height artefacts are a cantilever. It is sensitive to nano-scale surface topography variation, as
briefly presented. Besides, measurement uncertainty is evaluated for shown in Fig. 1, AFM head [8] has an AFM cantilever with a tiny tip
demonstrating an assurance in nano-scale step height calibration. driven by an Z piezo crystal, a Piezo linear motor introducing a
displacement in a direction of Z aixs and those Clamp piezos mounted at
2. Description of the metrological AFM separated positions inside he housing Tube. A larger approaching
movement either up or down moving along the z-axis is performed to
Commercially available AFMs have been widely employed to capture keep the cantilever in the same bending status. In this case, the AFM
surface topography with a nano-scale resolution for a number of head with its AFM tip cantilever keeping in a constant bending status is
different surface structures. Corresponding X, Y and Z surface used as a zero indicator to trace the testing surface dynamically during
the AFM tip scanning over the surface. Referring to an inchworm

https://doi.org/10.1016/j.measen.2021.100109

Available online 20 September 2021


S. Wang Measurement: Sensors 18 (2021) 100109

Fig. 3. The system in operation.


Fig. 1. Description of the inchworm moving for AFM head approaching the
testing surface.
metrological-AFM system.

2.2. Arrangement of the proposed system

A photograph of the metrological-atomic force microscope (M-AFM)


built in our centre is shown in Fig. 3. It is noted that system is located
inside a clean booth (class 100) environment to mainly provide a clean
environment to minimize possible dust containment onto he testing
sample. Also, there are an isolation table and an acoustical enclosure
taken into account keeping the system prevented from external me­
chanical vibration & sound effects in order to ensure the system working
environmental condition with less impacts to a high accurate measure­
ment and calibration.

3. Uncertainty analysis

In terms of the ISO guide to the expression of the uncertainty in


measurement (GUM) [10], the step height expanded measurement un­
Fig. 2. A schematic of a nano-scale platform.
certainty in the proposed system can be evaluated through considering
possible known source of errors. The step height (h) may be defined as
principle, the scanning Piezo linear motor is capable of adjusting AFM follows:
cantilever in AFM head gently and smoothly approaching the testing ( )
surface with a stepwise up and down scanning with a moving step-wise h= N
c cos θt
+ LAbbe + Ldp + Lmf *[1 − α(T − 20)] (1)
size less than 0.5 μm. A total Z movement range is up to 5 μm during 2nf0 cos θm
AFM head scanning along Z-axis to collect surface topography infor­
mation in the measurement. Where.
Fig. 2 shows a schematic diagram of a nano-scale platform produced
by SIOS Messtechmik GmbH [9]. The platform has a linear positioning N is the separation in interference fringes in the z axis of the two
range of 25 mm × 25 mm × 5 mm along X, Y and Z axes respectively. surfaces including step’s top & bottom sections on an area of interest
The platform is able to provide a minimum moving step in an order of on the step depth in terms of ISO 5436-1
0.1 nm and a moving linear displacement with an accuracy in order of c is the light speed in vacuum.
several nanometres. The AFM probe in use can be mounted above the n is the environmental air refractive index.
stage using the three (3) cylindrical holders made from Zerodur to f0 is the He–Ne laser frequency
ensure its thermal stability. Since the Zerodur material has a very low θm is laser incident angle with respect to laser reflective corner
thermal expansion coefficient (α is about 0.02 × 10− 6/◦ C), the three mirror surface
holder can stably keep all relevant measurement components, including θt is the residual deviated angle of the actual scanning axis with
AFM head, laser interferometers and mechanical parts, with minimum respect to and the measurement axis
drifts during the measurement in the system. There are 3 He–Ne laser LAbbe is a resulting Abbe offset error
linear displacement interferometers for directly determining each axis’s Ldp is a resulting error from the length of dead path in laser
linear displacement. Meanwhile, another two infrared optical auto­ interferometer
collimation angular sensors are used to synchronously correct Lmf is a resulting error from the thermal expansion about its metro­
nano-scale stage angular movement errors, such as the corner mirror logical frame
block’s yaw, pitch and roll angle deviation during AFM scanning with T is an averaged temperature during step height measurement
respect to the testing surface. Therefore, the stage can always move in αis the sample’s thermal expansion coefficient
parallel along X, Y and Z axes accordingly. In addition, it is worthy of
noting that the cantilever tip in AFM head is pre-adjusted to put it to be In terms of GUM, corresponding step height measurement uncer­
at the virtual interaction point of the three laser interferometers in order tainty budget could include contributions of type A and type B.
to make the AFM scanning direction in line with the Abbe principle. In The type A standard uncertainty is evaluated from a measurement
this case, there is a minimized Abbe offset error in the proposed repeatability to reflect the system measurement stability. Using a set of
21 step height measurements on a step height of 2000 nm, the resulting

2
S. Wang Measurement: Sensors 18 (2021) 100109

Table 1
Uncertainty budget for the step height calibration & measurement using LRM-AFM, number of measurements m = 21. Nominal values in use: laser frequency f0 =
474′ 1012 Hz, light speed c = 3′ 108 m/s, Vacuum air refractive index n = 1, Step height h, Abbe offset δ = 1 mm, dead path dp = 2 mm, misalignment angles qm = 6◦ and
qt = 7◦ , Averaged temperature T = 22.60 ◦ C.
i Quantity Xi Uncertainty u (xi) Probability Sensitivity coefficient Uncertainty contribution ui(h) Degrees of freedom
distribution ci (nm) vi

1 Interferometer data, N 0.0015 fringe R 316.4 (nm) 0.47 ∞


2 Vacuum frequency, f0 5.47 × 106 Hz R 2.11 × 10− 15
h (nm/ 1.15 × 10− 8h ∞
Hz)
3 Air refractive index, n 0.71 ppm N h (nm) 7.1 × 10− 7h ∞
4 Cosine error, θm 2739 ppm N 0.10h (nm) 2.74 × 10− 4h ∞
5 Cosine error, θt 3726 ppm N 0.12h (nm) 4.47 × 10− 4h ∞
6 Abbe error, LAbbe 0.56 nm N 1 0.56 ∞
7 Dead path,Ldp 0.12 nm R 1 0.12 ∞
8 Metrology frame expansion correction, 0.023 nm R 1 0.023 ∞
Lmf
− 7 − 9
9 Temperature, T 0.015 ◦ C R 5.0 × 10 h (nm/◦ C) 7.5 × 10 h ∞
10 Thermal coefficient α 1.2 × 10− 6/◦ C R 2.6h (nm.◦ C) 3.12 × 10− 6h ∞
11 Measurement repeatability, R 0.084 nm @2000 – 1 0.084 20
nm

Fig. 4. A typical result of a step height standard (a nominal value: 8 nm step


height) measured by the proposed system: 2D plot to show one of cross-section
perpendicular to step.

Fig. 5. A typical result of an undercut step height standard (a nominal step


type An uncertainty u(R) of the system is less than 0.084 nm, which height value: 1800 nm) measured by the proposed system: a 2D plot to show
mainly indicates random errors in the system. The relevant sources of one of cross-sections.
type B uncertainties in the proposed system can be given in Table 1.
As an error budget for any particular measurement may be calculated /
from appropriate summation of the estimated values of the individual 4

11
( )
veff = [uc (h)] u4i (h)/vi
contributions given above. Table 1 shows those uncertainty (3)
i=1
components. /( )
In accordance with GUM, corresponding combined standard uncer­ = 1.284 0.0844 ÷ 21 = ∞
tainty attributed to h can be calculated by the root sum squared of all From the t-distribution table, for a confidence level of approximately
errors in Table 1. 95%, the expanded measurement uncertainty is given by
From the t-distribution table, for a confidence level of approximately
95%, the expanded measurement uncertainty is given by U(h) = k ·√
uc ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
(h)
(4)
[ ]12 =2 × 0.746 2
+ ( 0 . 52 h )2

11
uc (h) = u2i (h) × h2 Therefore, the expanded measurement uncertainty of step height (h)
calibration in quadrature is as follows:
i=1

{( ) [( )2 U(h) = Q[1.5, h] nm, h in micron, estimated at a level of confidence of


= 0.472 + 0.562 + 0.122 + 0.0232 + 0.0842 + 1.15 × 10− 8
approximately 95% with a coverage factor k = 2.
( )2 ( )2 ( )2
+ 7.1 × 10− 7 + 2.74 × 10− 4 + 4.47 × 10− 4
( )2 ( )2 ] }12 4. Verification results of step height measurement
+ 7.5 × 10− 9 + 3.12 × 10− 6 × h2
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ In order to verify step height measurement accuracy in the proposed
= 0.7462 + (0.52h)2 (2) system, three step height standards have been utilized. The measure­
h in micron, the effective degree of freedom for uc (h) is given by ment results are described as follows.
A protrude step height standard with a nominal value of 8 nm was
measured. As shown in Fig. 4, its 2-dimensional (2D) AFM scanned re­
sults were obtained. Based on the step height calculation method

3
S. Wang Measurement: Sensors 18 (2021) 100109

defined in ISO 5436–1 [11], corresponding mean measured step height their surface topography measurements. It covers those applications in
was determined to be 8.6 nm with an expanded measurement uncer­ micro- & nano-scale structure measurements in the fields of MEMS,
tainty of 1.5 nm. In comparison of the measured value to its reference semiconductor, precision engineering and photonics industrial sectors.
value (8.5 ± 1.7) nm certificated by National Physical Laboratory (NPL,
UK), there is a deviation of 0.1 nm, which demonstrated that the step
height measured is favorably comparable to the reference value. Acknowledgments
An undercut step height standard artefact having a nominal step
height value of 1800 nm was calibrated by NPL and its reference value This research did not receive any specific grant from funding
was reported to be (1785.9 ± 3.8) nm. Corresponding 2D plot results agencies in the public, commercial, or not-for-profit sectors.
were obtained as shown in Fig. 5. The mean measured step height was
measured to be 1785.6 nm with an expanded measurement uncertainty
References
of 2.5 nm by the proposed system. The deviation of 0.3 nm from the
reference value is 0.3 nm. Similarly, taking into account the measure­ [1] G. Binng, C.F. Quate, C. Gerber, Atomic force microscope, Phys. Rev. Lett. 56 (9)
ment uncertainty of 3.8 nm from NPL, the measured value is also (1986) 930–933.
favorably comparable to the reference value. Therefore, the proposed [2] Gaoliang Dai, Helmut Wolff, Pohlenz Frank, Hans-Ulrich Danzebrink,
A metrological large range atomic force microscope with improved performance,
system is capable of conducting a nana-scale step height measurement & Rev. Sci. Instrum. 80 (4) (2009 Apr), 043702.
calibration. [3] J. Haycocks, K. Jackson, Traceable calibration of transfer standards for scanning
probe microscopy, Precis. Eng. 29 (2005) 168–175.
[4] J.A. Kim, J.W. Kim, B.C. Park, et al., Measurement of microscope calibration
5. Conclusions standards in nanometrology using a metrological atomic force microscope, Meas.
Sci. Technol. 17 (7) (2006) 1792–1800.
In this paper, a metrological atomic force microscope and its step [5] G.B. Picotto, M. Pisani, A sample scanning system with nanometric accuracy for
quantitative SPM measurements, Ultramicroscopy 86 (1–2) (2001) 247–254.
height measurement uncertainty have been described. The developed [6] V. Korpelainen, A. Lassila, Calibration of a commercial AFM: traceability for a
system is composed of an AFM for surface detection, a three-axis nano- coordinated system, Meas. Sci. Technol. 18 (2007) 395–403.
positioning stage for sample holding and scanning, control electronics & [7] S.H. Wang, G. Xu, S.L. Tan, Development of A Metrological atomic force micro­
scope for nano-scale standards calibration, in: Proceedings of SPIE, vols. 7155–17,
software, an anti-vibration isolated platform and acoustical enclosure.
2008.
In the experiment for verifying measurement uncertainty, step height [8] Danish micro engineering. http://www.dme-spm.com, March 2021 as it is in.
standards have been utilized to validate its nano-scale accuracy mea­ [9] SIOS Messtechmik GmbH, Am Vogelherd 46, D-98963 Ilmenau, Germany,
surement performance. Both experimental results and measurement http://www.sios.de, March 2021. as it is in.
[10] 1995 Guide to the Expression of Uncertainty in Measurement (Geneva: ISO).
uncertainty evaluation demonstrated that the system is able to provide a [11] ISO 5436-1, Geometrical Product Specifications (GPS) ─ Surface Texture: Profile
nano-scale measurement uncertainty in step height calibration. Because Method. Measurement Standards ─ Part 1: Material Measures, 2000, pp. 9–10.
three (3) He–Ne laser linear displacement interferometers are built in
along three (3) axes (X, Y and Z) in the system, the proposed Shihua Wang
metrological-AFM is then capable of providing a direct traceability of National Metrology Centre, Agency for Science, Technology and Research
step height measurement & calibration to disseminate a higher accuracy (A*STAR), 2 Fusionopolis Way, 138634, Singapore
to those commercial AFMs and high resolution 3D optical microscope for E-mail address: wang_shihua@nmc.a-star.edu.sg.

You might also like