On the suitability of alloys with a high molybdenum content

for the production of corrosion-resistant micro process equipment by diffusion bonding - Part 1 of 2 - Corrosion aspects, materials and corrosion and diffusion bonding tests

The wall thickness of micro-process engineering equipment is often less than one millimeter. This allows large amounts of heat from highly exothermic reactions to be transferred and processes to be operated continuously instead of in batches. As corrosion is a system property, laboratory tests do not always accurately reflect reality. Outstanding corrosion resistance is therefore essential. Microprocess engineering equipment is often constructed from a large number of microstructured sheet metal layers. They are joined by diffusion bonding at high temperatures. The parameters of temperature, duration and surface pressure must be optimized for each material, as the deformation differs.

Two materials with a high molybdenum content, Hastelloy B3 (2.4600) and Hastelloy BC-1 (2.4708), were examined with regard to corrosion in 70 % sulphuric acid at 100 °C for 1000 hours. In addition, diffusion bonding tests and tensile tests were carried out to determine the mechanical properties.

Introduction and motivation Corrosion aspects

Sulphuric acid is a common starting chemical in the chemical industry. Depending on concentration, impurities and oxygen content, its corrosion behavior can vary within wide limits. Up to a concentration of 65 % it has a reducing effect, above this it has an oxidizing effect [1]. This makes the selection of suitable materials, particularly in micro process engineering, difficult or subject to great uncertainty. Due to the very low material thicknesses, corrosion must be ruled out as far as possible in micro-process engineering equipment. Intergranular corrosion is often the cause of component failure as a result of sensitization through heat treatment, which occurs during diffusion bonding with a very low cooling rate.

Manufacturer specifications on removal rates in millimeters per year are irrelevant for micro process engineering. Corrosion tests for intercrystalline corrosion often used in practice, for example according to ASTM G28A or modified ASTM A262-B (Streicher test), are meaningless [2, 3]. On the one hand, this is due to the alloy-specific and very short test durations of these tests in order to quickly obtain a statement for the practitioner, but also to defined corrosion media and test temperatures that do not reflect the practical application.

Materials investigated to date

In the past, the Institute of Micro Process Engineering has already observed different service lives of micro apparatus in contact with hot sulphuric acid for different batches within the specified fluctuation range of alloying elements for Alloy 22 (2.4602) [4]. This underlines the difficulty in making statements about the suitability of a material.

It should also be noted that corrosion resistance is a system property that is influenced by a variety of factors (Fig. 1).

Fig. 1: Factors influencing corrosion resistance

Another problem is deviations in reaction kinetic data obtained from industrial applications. Since mixing processes are much more intensive in micro-process engineering, higher temperatures occur.

In the past, as part of an AiF project, systematic corrosion tests were carried out in concentrated and 70% sulphuric acid at 100 °C for 1000 h (six weeks) on four different high-alloy materials (Table 2) [5]. The material selection was based on the development history of the Hastelloy C family, which is presented in [6], and on suggestions of new developments by VDM Metals (Table 1) [7]. Their chromium content was between 20 and 26 %. The molybdenum content was up to 15 %.

Tab. 1: Composition of the materials

Tab. 2: Mass loss of the materials as delivered after six weeks of testing

The same tests were carried out in the heat-treated state (at 1100 °C, holding time 4 h with subsequent slow cooling in the diffusion bonding furnace) (Tab. 3).

Tab. 3: Mass loss of the materials after sensitization by heat treatment at 1100 °C and 4 h, after six weeks of testing

The results were not consistent throughout, especially for the heat-treated condition. Corrosion was assessed as too severe for micro-process engineering components for all materials. One reason for this is that the attack in the as-delivered condition was predominantly planar, while intergranular attack occurred in some cases after heat treatment. Mass losses are therefore of little significance here.

Manufacturing aspects of micro-process engineering equipment and diffusion bonding

The production of micro-process engineering equipment is complex. As face-centered cubic materials, high-alloy austenitic stainless steels or nickel-based alloys are difficult to machine due to their tough material behavior and cause high tool wear. Saws or milling cutters with finger cutters down to a few tenths of a millimeter in diameter are used.

Under certain circumstances, internal stresses are released in the sheet metal or the machining is accompanied by work hardening and leads to warping of thin sheets. Burrs on the microstructures must be removed before diffusion bonding in order to ensure vacuum tightness. Alternatively, iron-based materials can be microstructured by chemical etching.

Diffusion bonding takes place at approx. 80 % of the melting temperature of the materials, calculated in Kelvin. The process typically takes place under a high vacuum. This prevents oxidation and the components remain bright. However, the heat transfer takes place mainly through infrared radiation, which considerably reduces heating and cooling rates depending on the size of the furnace and component. At the set temperature, an external joining force is applied via stamps.

As a result, holding times in the range of hours not only lead to diffusion of atoms, but also to grain growth across the joining planes. The result is a monolithic body with mechanical properties that ideally correspond to those of a heat-treated workpiece.

The process is successfully used in tool mould construction to create cooling channels close to the cavity. A significant difference to micro process technology is that the steels used for this undergo a temperature-dependent polymorphic transformation, which considerably simplifies welding. The time-dependent grain growth during the holding time at high temperature is also equalized during cooling. These steels also have a much lower alloy content. Stainless austenitic steels or nickel-based alloys, on the other hand, form dense and oxygen-impermeable passive layers on the surface. They usually consist of chromium and nickel oxides that form spontaneously in the presence of atmospheric oxygen. They are only a few nanometers thick. They are transparent, but can be considerably strengthened chemically or by temperature, e.g. during rolling processes. Their thickness can then reach several hundred nanometers without this being visible on the surface [8].

Passive layers are responsible for the corrosion resistance of these materials, but seriously hinder diffusion bonding. For solid components such as injection molds, deformation during diffusion bonding is largely uncritical in terms of their function, as finishing is only carried out afterwards. In the case of micro-process engineering components, deformation of microchannels with cross-sections of a few hundred micrometers has a much greater effect on throughput and pressure loss. On the other hand, as the number of thin sheet metal layers for larger dimensions increases, 1. thickness tolerances of the sheets from rolling and 2. deformations from the leveling of the surface roughnesses increasingly add up.

Nevertheless, vacuum tightness must be ensured reliably and reproducibly as standard.

The deformation of micro-process equipment during diffusion bonding does not only depend on the effective joining surface. Material composition and grain size also have an influence on creep behavior at high temperatures. Therefore, the joining temperature, surface pressure and holding time must be optimized for each material.

High molybdenum materials

Haynes International provided coupons made of Hastelloy B3 (2.4600) and Hastelloy BC-1 (2.4708) with a thickness of 3 mm and a TIG weld seam for corrosion tests. The composition of both materials is shown in Table 4. It is noticeable that the molybdenum content of Hastelloy BC-1 is only slightly higher than that of Alloy 2120 MoN (2.4700), which was already investigated in the AiF project.

Tab. 4: Composition of Hastelloy B3 (2.4600) and Hastelloy BC-1 (2.4708)

Sheet material for diffusion bonding tests was also purchased from Haynes Int. As not all dimensions are available for these special materials, the thickness was 1.6 mm for Hastelloy B3 and 1 mm for Hastelloy BC-1.

Round material in various diameters for the production of tensile specimens after diffusion bonding was obtained from Zapp, Ratingen.

Experiments Corrosion tests

Corrosion tests were carried out analogous to the AiF project in concentrated and 70% sulphuric acid at 100 °C for 1000 h, both for the delivery condition and in the diffusion bonding furnace identically heat-treated at 1100 °C for 4 h with subsequent slow cooling (Tab. 5 and 6).

Tab. 5: Corrosion tests in 95-97 % sulphuric acid at 100 °C, 1000 h

Tab. 6: Corrosion tests in 70 % sulphuric acid at 100 °C, 1000 h

In general, it is noticeable that the percentage mass losses are less than 1%, with the exception of the heat-treated state of Hastelloy B3 in 70% sulfuric acid. The mass losses in 70% sulphuric acid are also not significantly greater than in concentrated sulphuric acid, as had been expected. Lower values in the heat-treated condition are attributed to errors in the area of measurement inaccuracy.

Diffusion bonding tests

Preliminary tests on both materials were used to determine the deformation behavior and to assess the microstructure. For this purpose, ten 20 x 20 mm sheet layers were cut out using a laser cutting machine and diffusion welded at 1100, 1150 and 1200 °C respectively. It should be noted that the sample height varied due to the different sheet thicknesses. Due to the low initial height and the low absolute deformation, the measurement error when calculating the percentage deformation is relatively large. The surface pressures were reduced for higher joining temperatures. It was found that the reduction more than compensated for the increase in flow velocity with temperature, so that the deformations even tended to decrease for higher temperatures (Table 7). Continued on p. 749

Tab. 7: Preliminary tests for diffusion bonding Hastelloy BC-1 and B3 from ten sheet layers each

It should also be noted that the deformation also depends on the geometry, in particular the aspect ratio (ratio of height to diameter).

The joint quality was assessed on cross-sections of the samples. Experience has shown that a deformation of at least 3 % is helpful to compensate for differences in sheet thickness and to level out surface roughness. This must always be assessed in conjunction with the absolute sample height, number of layers and quality of the sheets, etc.

The microstructures showed no grain growth across the joining planes below 1200 °C joining temperature. As the deformations were also very small, the joining parameters T = 1200 °C, p = 10 MPa, t = 4 h were selected to produce diffusion bonding specimens from which tensile specimens could be produced to determine the mechanical properties (Fig. 2).

Fig. 2: Dimensions of the diffusion bonding specimens for the production of tensile specimens

Although the aspect ratio of approx. 1.5 was significantly higher than in the previous tests and the surface pressure was increased by two thirds from 6 to 10 MPa compared to the previous test at the same temperature, the deformation of these diffusion bonding specimens was 2.03 % for Hastelloy B3 and 2.06 % for Hastelloy BC-1. This is still at the lower limit of the necessary deformation. Further optimization is necessary here.

Tensile tests as well as discussion and conclusions will follow in Part 2 of Galvanotechnik 7/2024.

Literature

[1]DECHEMA material tables, properties of sulphuric acid, online database at https://corrosionhandbook.de/,(fee-based access) or G. Kreysa; M. Schütze: Corrosion Handbook, vol. 11, (2008), 2nd completely revisededition, Wiley-VCH, Weinheim.
[2]Corrosion test for intergranular corrosion ASTM G28, 2015, see https://www.astm.org/Standards/G28.htm, ASTM International, West Conshohocken, PA, USA, last accessed on 21.04.2021
[3]Corrosion test for intergranular corrosion ASTM A262, 2015, see at https://www.astm.org/Standards/A262.htm, ASTM International, West Conshohocken, PA, USA, last accessed on 21.04.2021
[4]Data sheet and properties of Alloy 22, VDM Metals, data sheet as of November 2020, see at https://vdm-metals.com/de/alloy22/, last accessed on 21.04.2021
[5]Final report AiF-Project No. 18034 N, Investigations to improve the corrosion resistance of micro-process engineering components for aggressive chemical process media, duration: 01.02.2014-30.09.2016, see at http://www.imvt.kit.edu/downloads/AiF_Abschlussbericht_20170117_gie.pdf,last accessed on 21.04.2021
[6]D.C. Agarwal; W.R. Herda: The "C" family of Ni-Cr-Mo alloys' partnership with the Chemical Process Industry: the last 70 years, Mater. Corros. vol. 48, (1997), 542-548
[7]H. Alves; R. Behrens; L. Paul: Evolution of Nickel Base Alloys - Modification to Traditional Alloys for Specific Application, Corrosion 2014, March 9-13, 2014, paper No. 4317, San Antonio, Texas, USA, NACE International
[8]T. Gietzelt; V. Toth; T. Weingärtner: Impacts of Layout, Surface Condition and Alloying Elements on Diffusion Welding of Micro Process Devices, Mat.-wiss. u. Werkstofftech., vol. 9, (2019), 1070-1084, doi: 10.1002/mawe.201800197