Expert Talk

Quantitative evaluation of phase fractions in duplex stainless steels

Jan Y Jonsson

Senior Specialist Physical Metallurgy
Outokumpu

Jan Y Jonsson outlines the main methods used for determining the phase fractions in duplex stainless steels.

What is duplex stainless steel – and why are the phase fractions important? 

Duplex stainless steel grades have a ferritic-austenitic microstructure, with a phase balance of approximately 50% ferrite and 50% austenite. They combine many of the beneficial properties of ferritic and austenitic stainless steels.

The duplex microstructure contributes to the high mechanical strength and high resistance to stress corrosion cracking. Knowing the fraction of each phase present in the steel is vital in establishing that the material will have the desired properties. 

 

The main methods 

There are three main methods for determining the phase fraction in a duplex stainless steel:

  • Manual Point Counting (MPC) to ASTM E562 – this is widely used and often required 
  • Image Analysis (IA) to ASTM E1245 – generally the better option
  • Magnetic methods – can be easy to use but subject to many influencing factors

 

Manual Point Counting (MPC) 

With the manual point counting (MPC) method a point grid of between 16 to 400 points is distributed on the microstructure of examined sample as shown in Figure 1. The number of hits measured in each phase is then recorded. To get good statistics this method is normally time-consuming but is a good choice for even size phase distributions i.e. good for base material but not weldments and secondary austenite). It does though offer little or no information on overall distributions (surface to surface variations) and it is subjective. It should be emphasized that using an active decision for setting value of grid point improves the accuracy for this method. Using semi- automatic tools also makes this method faster.

Figure 1: MPC grid

 

Image analysis (IA)

Image analysis uses computer software to analyse the microstructure on a pixel basis– see Figure 2. One can use either a manual or automatic set threshold and the method has the benefit of being fast. It is a good choice for most distributions and provides information on the overall distribution. Unlike MPC, IA is objective rather than subjective. But its success does rely on the quality of the etching.

image analysis 
Figure 2: image analysis

 

Magnetic method – Ferrite scope 

The Ferrite scope – see Figure 3 - uses magnetic permeability, rather than saturation magnetism, to determine the volume analysis of the ferrite content. It is a very fast method using a hand-held probe. It may provide information on overall distributions but is difficult to apply for very small volumes. It needs good calibration and can then provide objective information. 

Figure 3: Ferrite scope
Figure 3: Ferrite scope

 

Issues with deep etching 

Deep etching of a sample can cause issues that result in either over- or under-estimation of the ferrite phase as illustrated in Figure 4.

Figure 4: issues with deep etching

 

Etching of white phase 

Etching of the white phase in LOM (light optical microscopy) may give an over-estimation of the dark phase (ferrite in the example shown in Figure 5) if pronounced.

Figure 5: issues with white phase etching

 

The effect of contrast in etching for IA

To have enough contrast between ferrite and austenite is very important for obtaining adequate results with IA. In Figure 6 an image with narrow contrast is shown and where it can be seen that “phase boundary”(6%) also wrongly include parts of the ferritic area. In Figure 7 on the other hand, with high contrast, a phase boundary area of 4% is detected and no obvious faults in the threshold can be seen. 

Figure 6: Low contrast - 4mm thick 1.4410, NaOH. 47% ferrite, 6% phase boundary
 
Figure 7: High contrast - 4mm thick 1.4410, NaOH. 49% ferrite, 4% phase boundary 

 

Phase boundary area effects in IA and MPC

As shown in Figures 8 and 9, there can be phase boundary area effects in both IA and MPC. In Figure 8 the low contrast show phase boundary to be obviously wrong (not only boundaries) and for MPC the decision for setting phase will be tougher as more boundaries exist in each image. In Figure 9 the opposite situation exists with the very high contrast gives perfect thresholding for IA as  well as the decision in MPC becomes easier with less boundaries in the image field. 

Figure 8: 1mm thick 1.4410 NaOH 64% ferrite, 9% phase boundary 
Figure 9: 30mm thick 1.4362 Beraha II. 45% ferrite, 2% phase boundary 

 

Comparison of EBSD and IA 

EBSD (electron backscatter diffraction) in an SEM (scanning electron microscope) is a useful reference technique for phase fraction measurement. One example is shown in Figure 10 where IA only over-estimates the ferrite content by 2% in a relatively tough structure with different structure sizes.

Figure 10: Comparison of EBSD and IA

 

Comparison of EBSD and IA at high magnification

At higher magnification the over-estimation of the ferrite content by IA becomes more pronounced. In the case shown in Figure 11, EBSD indicates a ferrite content of 51% and IA 56%. 

Figure 11: Comparison of EBSD and IA at high magnification

 

Comparison of IA and the Ferrite scope 

Figure 12 shows a comparison, for different materials, of the ferrite content measured by IA and Ferrite scope. In some cases the Ferrite scope produces an under-estimate, in others it over-estimates. Optimally it has been shown that a ferrite scope needs to be calibrated for alloy and product. 

Figure 12: Ferrite % by IA on the y-axis and Ferrite scope on the x-axis 

 

Comparison of IA, MPC and EBSD 

Figure 13 shows that there is generally a reasonable correlation between the results of EBSD, IA and MPC for good etching. However, bad etching will result in both IA and MPC over-estimating the ferrite content. 

Figure 13: comparison of ferrite percentage as measured by EBSD, IA and MPC

 

Different types of measured sections and thickness correction for the Ferrite scope 

In Figure 14 we see the significant variation in the reading produced by the Ferrite scope depending on the type of cross-section that is measured. While the second and third group of bar graphs show how the reading can be corrected for thickness. 

Figure 14: Impact of section and thickness correction on the Ferrite scope reading 

 

Conclusions

  • IA is more accurate and faster than MPC. It includes distribution if needed and is the best option for evaluating small amounts and the best option for evaluating welds. The phase boundary area can be used to give procedural accuracy (etching and imaging)
  • MPC is the best option when etching does not give adequate contrast for the IA threshold. The method can be significantly improved by using semi-automatic tools and also by taking active decisions when setting value for grid points.
  • The Ferrite scope is faster than MPC and IA. But it does need to be calibrated for alloy, measured cross-section and thickness to give results within ±5% accuracy.
  • EBSD requires excellent polishing