**2. Materials**

Two laboratory heats of 15 kg, coded here as CR1 and CR3, were produced in a vacuum induction furnace under an inert atmosphere. Alloying elements were added in sequence to pure (>99.9%) electrolytic iron. Carbon deoxidation was performed and an analysis of C, S, N, and O was made on line during the final adjustment of the composition. Samples of 40 mm thick plates were hot rolled to a final thickness of 3 mm in several passes finishing at 900 ◦C. The 3 mm thick strips were then cold rolled to blanks with a thickness of 1.3 mm. Table 2 shows the chemical compositions of the experimental steels determined by optical emission spectroscopy (ARL 4460, Thermo Fisher Scientific, Lausanne, Switzerland). Critical transformation temperatures *Ms* and *Ac*3, determined by high resolution dilatometry are also included in Table 2. It also shows the times required for completing bainitic transformation, determined by dilatometric analyses at temperatures at and above *Ms*. If isothermal treatment takes place at a temperature at or below *Ms*, athermal martensite forms before the isothermal transformation starts. For more information about design, processing, and properties of isothermally treated CR1 and CR3 steels, see Caballero et al. [23]. In addition, the composition range of the 22MnB5 reference steel used in final hat-profile pressing is given in Table 2.


**Table 2.** Chemical compositions (wt %) of experimental CFB steels and the reference 22MnB5 (B5) steel. Experimentally measured *Ms* and *AC*<sup>3</sup> temperatures as well as bainite formation time (*tBf*) at select isothermal holding temperatures (*TB*) are also included.

## **3. Methods**

#### *3.1. Gleeble Thermal Simulation*

Experiments simulating press hardening conditions in respect of temperature—time cycles were carried out using a Gleeble 1500 simulator (Dynamic Systems Inc., Postenkill, NY, USA). Flat specimens with dimensions 1.3 × <sup>10</sup> × 70 mm<sup>3</sup> were subjected to thermal cycles that produced a uniform heat-affected zone of about 20 mm in width in the center of the samples. Thermal cycles were designed to simulate two industrial processing routes in which the following sequence of steps is used: (i) austenitization; (ii) forming (at a specified temperature); (iii) quenching to a specified temperature to simulate either an austempering treatment above the *Ms* temperature or a QP process below the *Ms* temperature (resulting in a certain amount of martensite formation); (iv) cooling the austempered samples or heating the QP samples to a specified temperature both above *Ms* respectively, followed by isothermal holding to facilitate transformation of, some or all of, the balance untransformed austenite

into a very fine bainitic structure and; v) cooling to room temperature (after complete or partial transformation to bainite). The Gleeble simulations performed in this work however, did not include blank deformation, i.e., step (ii) in the sequence above.

All specimens were first reheated at 5 ◦C/s to 930 ◦C and held for 60 s before cooling at 20 ◦C/s to 770 ◦C. Two different sequences of heat treatment were then performed, directly without delay:


Following the heat treatments, samples were cooled to room temperature at 5 ◦C /s. Referring to the QP process the athermal martensite fractions formed at temperatures of *Ms* − 10 ◦C and *Ms* − 20 ◦C were estimated to be ~10 and 20 vol %, respectively using the Koistinen-Marburger equation [24].
