Logo PTB
Production sequence of Si-spheres and interferometrical determination of the sphere volume

AFM Linewidth Metrology

Working Group 5.23

Developments of high resolution CD/3D AFMs


Progressive miniaturisation in nanomanufacturing is delivering ever smaller nanodevices with more complicated 3D structure and material composition. For instance, the pace of feature size reduction in the semiconductor industry is now approaching technology nodes below 22 nm. For quality control and process monitoring, metrology of 3D structures of both silicon wafers and photomasks is required with a desired uncertainty of less than 1 nm. True 3D metrology of these structures including form (width, height, edge slope, corner rounding and footing) and line edge/width roughness is one of the most challenging metrology tasks today, especially the structure width which is usually referred to as the critical dimension (CD) in semiconductor industry.

Several techniques are available today for performing critical dimension (CD) metrology, for instance, the optical scatterometry with visible and shorter wavelength (EUV, x-ray), critical dimension scanning electron microscopy (CD-SEM) and 3D atomic force microscopy (3D-AFM). Each method has its advantages on one hand but disadvantages on the other hand. Scatterometry is fast and can be applied for inline measurements, however, it needs advanced mathematical tools in modelling and data evaluation for solving its inverse problem, where the geometrical 3D shape of the structures is required as the prior knowledge. CD-SEM is widely used for CD measurement for both wafers and photomasks. A resolution down to 1.5 nm and a reproducibility in the sub-nanometre range can be achieved using electron energies smaller than 1 keV. However, it has to be performed in vacuum, lacks 3D measurement capability, and is sometimes destructive. AFM technique allows direct measurements of the 3D shape of nano structures with both a high lateral and vertical resolution (almost) non-destructively. Compared to other methods, it is the least model-dependent technique due to its relative straight forward measurement principle and the relatively better understanding of the tip sample interaction mechanism. Despites of low through put, CD/3D AFM is becoming one of the most applied techniques for 3D nanometrology and is widely applied in various industries.


CD/3D-AFM based on the Nanostation 300

A photo of the 3D-AFM developed by PTB is shown in figure 1. The system is based on a scanning head design. This affords the advantage that large samples such as wafers and photomasks can be held stationary during measurements while the tip is scanned.  A commercial 6-axis nano positioning stage (type P-561K034, Physik Instrumente GmbH) is applied for moving the AFM head. The nano positioning stage is a flexure hinge stage driven by piezo actuators, and its positions and angles are measured and servo-controlled using high resolution capacitive sensors. The displacement scale factors of the 3D-AFM are calibrated using a set of step height and lateral standards that were measured by the PTB metrological large range AFM.  The measurement range of the PTB 3D-AFM is 15 μm × 45 μm × 45 μm (x, y, z). An air bearing stage with a travel range of 550 mm x 300 mm is used for coarse positioning of the sample. To reduce vibration noise, the air bearing is deactivated so that the motion table is set down prior to performing AFM measurements. 


Figure 1: Photo of the CD/3D-AFM based on the Nanostation 300


CD-AFM and Tilting-AFM measurement modes

Two different AFM techniques, the CD-AFM and the tilting-AFM, have been developed in our instrument. The CD-AFM uses flared tips. Such tips have an extended geometry near its free end which enables the probing of steep and even undercut sidewalls, as shown in figure 2(a). The CD-AFM technique has advantages of measuring both the left and right sidewalls of nanostructures in only one measurement. However, it has disadvantages such as the relatively large tip geometry (typically tens to hundreds of nanometres) which limit its spatial resolution for measuring very dense structure patterns, and the complicated tip shape which makes the tip characterisation more difficult.

To extend the measurement capability of the CD-AFM, the tilting-AFM technique has been developed and implemented recently. The advantage of the tilting-AFM lies in its capability of applying very fine conical tips, for instance, super sharp silicon tips with a radius down to 2 nm, which offers better capability for measurement of dense structure patterns as well as for the corner rounding and footing of structures. But the tilting-AFM has its own limitations, too. In order to make the sidewalls of steep structures measurable, the AFM tip has to be tilted with respect to the structure by a certain angle, as shown in figure 2(b). As only one side of the structure is measurable at one tilted setup, either the AFM tip or the sample must be rotated so that the opposite side of the structure becomes measurable. Consequently, the images obtained at different tilting views must be stitched together to determine the CD, where the stitching error will strongly influence the measurement accuracy.

we apply both techniques in a complementary manner, adding their strengths and overcoming their limitations, and thus reducing the measurement uncertainty and expanding the metrology versatility.


Figure 2: Principle of the CD-AFM (a) and tilting-AFM (b) applied in the measurement


Vector Approaching Probing (VAP) technique

In order to minimize tip wear and enhance measurement flexibility, a so-called “vector approaching probing” (VAP) method is implemented in the 3D-AFM.  The structure is measured point by point using the VAP method. At each measurement point, the tip is moved towards the surface until the desired tip-sample interaction is detected and then immediately withdrawn from the surface. The probing direction is usually set to be normal to the surface for achieving highest probing sensitivity, however, other probing directions are also allowed. The VAP method is capable of measuring vertical or even reentrant sidewalls. 

 A VAP measurement process consists of three phases as shown in figure blow. The first is the approach phase. In this phase, the tip is moved towards the surface at a desired probing velocity vp (for instance 500 nm/s ~ 1000 nm/s) and the tip sample interaction is monitored in real time. The first phase ends once the tip sample interaction reaches a pre-defined status ST1. The second is the probing phase. In this phase, the tip is moved towards the surface until another pre-defined status ST2 is reached. Then the tip is withdrawn immediately from the surface until the tip sample interaction reaches again the status ST1. During the second phase, the x, y and z position of the nano-positioning stage and the probe signals of sample interaction are recorded in real time and stored in the buffer of the DSP system. The third phase is positioning for the next measurement point. The time needed for a complete VAP step is approximately 2dt-s/vp, where dt-s is the initial tip sample separation before the VAP starts. 

An example of a typical probing curve measured using the VAP method is shown in the figure 3. The curve is recorded measuring a horizontal surface with a tip probing in the vertical direction using the vertical oscillation mode.  All three probing phases are illustrated. The free oscillation amplitude of the tip Af is about 12 nm. Before the VAP process starts, the tip has a tip-sample distance of about 50 nm – shown as the point S0. After the process starts, the tip approaches the surface until it reaches the point P0. During this period, the oscillation amplitude Av changes little, because the tip is far from the surface and the tip-sample interaction force is almost zero. At the position P0, the tip-sample interaction becomes significant and Av starts to decrease. The approach phase ends at the point S1 where Av = AST1, and the probing phase is entered. During the probing phase, the tip is moved closer to the surface and Av continues to decrease. The motion is stopped at the point S2 where Av = AST2, and then the tip is withdrawn from the surface until the point S1 is reached again. In the third phase of the VAP method, the tip is moved back to its original position through the curve S1P0S0. After the probing curve is taken, a line is fitted to the data points between S1 and S2.  The surface position can then be calculated as the point P1 (Av = As) on the fitted line.  In the designed software, the values of AST1, AST2, and As are adjustable. 


 Figure 3: Principle of the “vector approach probing” is shown in (a) and a typical probing curve is shown in (b).


In order to achieve optimal probing sensitivity, the tip should approach and probe the sample along a direction normal to its surface. However, this requires prior knowledge about the form and the position of the structure. Unfortunately, at least for the initial profile, such prior knowledge is not available.  To overcome this lack of prior knowledge, a three step process for measurement of the first profile was developed. First, the structure is measured using the VAP method with a fixed probing direction – typically vertical. Second, as shown in figure 4, the measured profile is fitted to a number of different segments.  The measurement is then configured for each segment in terms of how many measurement points should be taken, which measurement mode should be applied, etc.  Third, the structure is then measured again using a probing direction normal to the surface. This is referred to as the 3D probing strategy.

When subsequent profiles are measured, the previous profile can be used as the prior knowledge from which to generate the 3D probing strategy for each new profile. This method is quite effective since adjacent scan lines over many relevant structures are often very similar.  Additionally, since the most recent data are used to generate the new strategy, this method can adapt well both to drift and the incremental differences in the actual profile at each location.


Figure 4: 3D Measurement strategy. The labels in the figure are: left base (LB), left bottom corner (LBC), left sidewall (LS), left top corner (LTC), top (T), right top corner (RTC), right sidewall (RS), right bottom corner (RBC) and right base (RB).


The VAP method has several major advantages. First, it offers more flexibility in measurements than conventional AFM scanning. The surface is measured point by point using the VAP method, which allows flexible definitions of the measurement points. The part of the structure which is the most important for any given measurand can be probed in more detail than less important regions.  To measure the middle CD of a structure, for example, the sidewalls can be measured with higher pixel density than on the upper and lower surfaces. In this way, the total number of measurement points can be greatly reduced.  This in turn reduces the measurement time and minimizes the tip wear.  Additionally, the probing direction, probing velocity, and probing mode at each measurement point can be configured independently for optimum performance at each position. The measured coordinate at each point is also evaluated from a full tip-sample interaction curve rather than coordinates of a single point. This entire tip-sample interaction curve provides much more information for data processing. Finally, in contrast with conventional AFM scanning in which the tip is continuously interacting with the surface, the tip-sample interaction time is much lower when using the VAP method – which further reduces tip wear. 


Typical measurement performances

In this section, some selected measurement results taken on an IVPS 100 sample, a SCCDRM sample, and a PTB photomask are described to demonstrate the typical measurement performance of the developed 3D-AFM. 

To demonstrate the imaging capability of the 3D-AFM, an image of a sample with GaN nanorods is shown in figure 5(a), along with a cross sectional profile, shown in figure 5(b), taken from the image at the location indicated.


Fig.5  An AFM image measured on a sample with GaN nanorods is show in (a) with a cross sectional profile at the marked position in (b).


An IVPS 100 type sample supplied by the company Team Nanotech has been measured to test the capability of the 3D-AFM in measuring the 3D form of structures. The sample contains silicon lines with vertical, parallel sidewalls (111-crystal planes).  The features have a nominal depth of 900 nm and a nominal sidewall angle 90°. In this investigation, a flared AFM tip of the CDR130 type was used in the vertical oscillation mode. Since the feature depth is greater than the effective length of the flare tip, a software function called a “virtual cut-off plane” was applied.  This function, which was developed to protect the tip in circumstances like this, limits the tip to measure only at points above the cut-off plane. To test measurement stability, five profiles were taken at the same position with the slow scan axis disabled.  These measured profiles are shown in figure 6, with three insets showing the details of both sidewalls and the top region. The profiles shown are the raw measurement data, including the dilation effects due to tip shape, without any smoothing or filtering. It can be seen from these results that the instrument is capable of resolving the 3D form of the structure with a repeatability of better than 1 nm (p-v).


Fig.6 Measured profile of a IVPS 100 sample with the insets showing the details at the marked areas. Five profiles repeatedly measured are shown in the figure.


Development of a new Low-Noise 3D/CD-AFM

(This part is under construction)

Two strategies applied in traceable calibration of the geometry of the reference nano structure based on its transmission microscopic images, (a) via silicon crystal lattice constant and (b) via metrological AFM

For more details of this research task, please refer to some selected publications listed below:

[1] Gaoliang Dai et al.  Gaoliang Dai, Ludger Koenders, Jens Fluegge, Harald Bosse, “Two approaches for realizing traceability in nanoscale dimensional metrology,” Opt. Eng. 55(9), 091407 (2016), doi: 10.1117/1.OE.55.9.091407.
[2] Gaoliang Dai et al. Reference nano-dimensional metrology by scanning transmission electron microscopy, Meas. Sci. Technol. 24 (2013) 085001
[3] Gaoliang Dai et al. Development and characterisation of a new line width reference material, Meas. Sci. Technol. 26 (2015) 115006
[4] Gaoliang Dai et al. Comparison of line width calibration using critical dimension atomic force microscopes between PTB and NIST, Meas. Sci. Technol. 28 (2017) 065010