Poster - Optimisation of surface topography characterisation for metal AM using CSI - Final

Full text

Manufacturing Metrology Team
Fig. 2. CSI measurements of metal AM surfaces: (a) S1, (b) S2, (c) S3, and (d) S4. The 1.4× objecve lens (1×
zoom) was used for (a), the 5.5× objecve lens (0.5× zoom) was used for (b), the 5.5× objecve lens (1×
zoom) was used for (c) and (d)
Fig. 3. Eects of spectral ltering and fringe analysis methods on data coverage. 5.5× objecve lens
was used. The data coverage is ploed as a funcon of a) Sq and b) Sdq
OPTIMISATION OF SURFACE TOPOGRAPHY CHARACTERISATION FOR METAL ADDITIVE
MANUFACTURING USING COHERENCE SCANNING INTERFEROMETRY
Carlos Gómez1, Rong Su1, Adam Thompson1, Jack DiSciacca2, Simon Lawes1 and Richard Leach1
1Manufacturing Metrology Team, University of Nongham, NG7 2RD, UK 2Zygo Corporaon, Middleeld, CT 06455 USA
Table 1. Surface texture parameters for the test cases. An S-lter and an L-lter were applied to remove
high frequency noise and long scale waviness/form
Fig. 4. Eects of signal oversampling on data
coverage. 5.5× objecve lens, narrow bandwidth
spectrum and coherence prole fringe analysis
method were used
Fig. 6. Eects of signal oversampling on data coverage for S4. 5.5× objecve lens, narrow bandwidth spec-
trum and coherence prole fringe analysis method were used. (a) Eect of camera exposure me; (b) meas-
urement without signal oversampling (0.5× zoom); and (c) result with 8× signal oversampling (0.5× zoom)
Fig. 5. Comparison between HDR and signal
oversampling and a measurement performed
without using advanced funcons. The data
coverage is ploed as a funcon of Sdq
Surfaces Sq /µm Sdq
S1 LPBF Al-Si-10Mg cube, top surface 19 ± 2 0.6 ± 0.1
S2 LPBF Al-Si-10Mg cube, side surface 20 ± 3 1.0 ± 0.2
S3 LPBF Ti-6Al-4V cube, top surface 22 ± 3 1.1 ± 0.1
S4 EBPBF Ti-6Al-4V rectangular prism, top surface 34 ± 2 1.7 ± 0.2
Summary and outlook
 The eects of the advanced measurement funcons on the measurements of several
typical AM surfaces have been demonstrated. Results show that the CSI technique is fea-
sible for measuring surface topography in metal AM
 Increasing the signal oversampling factor in combinaon with the use of a narrow band-
width source spectrum will maximise data coverage without sacricing measurement ar-
ea, but measurement me may be compromised. Recommendaons are provided for
the opmisaon of measurements on metal AM surfaces in terms of me, measurement
area and data coverage
 This study also presents insight into areas of interest for future rigorous examinaon,
such as measurement noise and further development of guidelines for the measurement
of metal AM surfaces
Movaon
 Metal AM surfaces are rough and can be challenging to measure, due to the presence of
complex features, which include high slopes, step like recesses and protuberances, and
local variaons in reectance
 Coherence scanning interferometry (CSI) is a non-contact measurement method that us-
es a broadband light source and interference to measure surface topography and object
geometry, originally designed for measuring smooth surfaces
 Recent progress in the development of the CSI technique allows a signicantly enhanced
detecon sensivity through advanced measurement funcons, such as ltering of the
source spectrum, high dynamic range (HDR) lighng levels, adjustable number of camera
acquisions over each interference fringe (i.e. oversampling) and sophiscated topogra-
phy reconstrucon algorithms
Method
A ZYGO NewView™ 8300 CSI system was used for this study. The experimental design co-
vers the following aspects:
 Four common metal AM surfaces made from dierent materials and processes (laser
powder bed fusion (LPBF), electron beam powder bed fusion (EBPBF))
 A series of measurements performed by using a combinaon of three objecve lenses
and two opcal zoom factors, two spectral lters, two fringe analysis methods, ve
sengs of signal oversampling and two HDR lighng levels. Topographic measurements
are described through areal surface texture parameters Sq and Sdq and are analysed for
data coverage, measurement me and area
Fig. 1. CSI data acquision (data rates are more than a million surface height points per second)
Objecves
 Demonstrate the feasibility of using CSI for characterising metal AM surfaces
 Evaluate the eecveness of relevant CSI measurement sengs
 Provide recommendaons for the opmisaon of measurements on metal AM surfaces
using CSI
CSI measurements of metal AM surfaces Eects of measurement funcons and sengs
S1/S2 S3 S4
Source spectrum ltering Narrow (40 nm BP lter)
Fringe analysis method Coherence prole
Zoom lens 1×
Objecve lens 1.4×* 5.5× 5.5×
Oversampling factor -- -- or 4* 2 or 8*
* For data coverage > 99 %
Table 2. Measurement opmisaon for metal AM surfaces (95% data coverage by default)