Strengthening mechanisms in Monel K500 alloyed with Al and Ti

Monel alloys containing 63Ni–30Cu (wt%) are often used in applications requiring simultaneously high strength and corrosion resistance. Additions of Ti, Al and C to Monel K500 lead to formation of TiC, Ni3Al and Ni3Ti particles, which provide precipitation strengthening effect following heat treatment. The traditional heat treatment schedule includes solution annealing above 1000 °C and aging in the 400–600 °C temperature range. However, no correlation exists between the alloy composition and the heat treatment schedule (holding temperature and time) required to obtain the optimum microstructure and mechanical properties. This may result in excessive alloying, energy loses during heat treatment, and higher product costs. In this work, we investigate the effect of solution annealing part of heat treatment schedule on microstructure (particularly, particle precipitation and grain growth), hardness and strength. For hot rolled samples, solution annealing followed by aging was shown to result in lower strength compared to aging without annealing. The analysis of strengthening mechanisms carried out utilising our theory for calculation of solute atom concentrations has explained the strength variation with heat treatment and has shown (i) a lower strength after annealing and aging being related to dissolution of fine (< 20 nm) TiC particles, (ii) Ti- and Al-rich precipitates to provide a larger strengthening than Ti and Al solute atoms, (iii) Al to be a more effective strengthening agent than Ti, and (iv) the majority of Al to remain in solution for both processing schedules, this indicates potential for mechanical properties improvement via optimisation of the heat treatment schedule aiming to generate more Al-rich precipitates.


Introduction
Ni-base alloys sustain high strength, ductility and toughness in a wide temperature range from absolute zero to about 1200°C. This is attributed to the austenitic matrix with face centred cubic crystal structure (which favours development of complex twin/dislocation sub-structure) and a high solubility of alloying elements in solid solution (which may result in broad size distributions of precipitates when the solubility decreases with a temperature decrease). Major additions of copper (28)(29)(30)(31)(32)(33)(34) wt% Cu) form the basis for Monel alloys family and provide solid solution strengthening and improved corrosion resistance, particularly in sea water, acidic and alkaline media, chloride solutions, and in many oxidizing and reducing gas environments [1][2][3][4][5][6]. Additions of Al, Ti and C to the Ni-Cu matrix followed by agehardened heat treatment help to improve strength via precipitation of c 0 -Ni 3 AlX (where X can be Cu, Mn, Ti or Si), Ni 3 Ti, NiFe 3 (AlFe) and TiC [7]. Due to their excellent properties, NiCu alloys are widely used for (i) manufacturing of machine components (propeller shafts, valves, fixtures and fasteners, evaporators, heat exchangers, blades of turbine and diffusers in steam jet ejectors) in chemical, petrochemical and nuclear industries [8,9], marine [10][11][12][13][14][15] and high temperature environments (gas turbines, missiles, aerospace) [9,16,17]; (ii) electrodes in fuel cells [18]; (iii) hydrogen generation [19][20][21], and (iv) corrosion resisting coatings [22,23]. In terms of processing technology, NiCu alloys can be rolled and forged [24], welded [25] and frequently subjected to agehardening heat treatment [26]. Our previous results have shown a possibility of wire arc additive manufacturing of NiCu alloys [27,28].
Traditional industrial practice, aligned in particular with the defence standard DEF STAN 02-771 ''Requirements, procedure and inspection for weld repair of copper alloy and nickel alloy castings'', stipulates a heat treatment schedule that includes solution annealing at elevated temperatures ([ 1000°C) followed by one or two cycles of age hardening at moderate temperatures (400-600°C). As a result of high temperature annealing, coarse particles are expected to dissolve and the matrix to become supersaturated with microalloying elements. This may facilitate precipitation of finer particle sizes during low temperature holding. High number density of nano-sized particles can provide significant precipitation strengthening effect [7,29,30]. However, the dislocation annihilation at high temperature may slow down pipe diffusion during aging, leading to increased particle nucleation times. This would require longer aging times and increase energy consumption. Decreased grain size strengthening (due to grain coarsening) and dislocation strengthening (due to a decrease in their density) may compensate for the increased precipitation strengthening effect, and the resultant strength will be low. Therefore, the effect of high temperature annealing should be clarified. In this paper, we characterised three conditions of Monel K500: as-received commercially hot rolled, laboratory aged without annealing, and aged after annealing. Annealed and aged samples exhibited lower hardness and compression yield stress than those that were aged without prior annealing. Thorough microstructural characterisation on scanning and transmission electron microscopes supported discussion on strengthening mechanisms operating in the studied alloy. (wt/at%) was used in this study. The rod was cut into disks perpendicular to its axis, and the disks were heat treated using two schedules: 1-heating at 10°C per minute to 610°C, holding at this temperature for 8 h followed by air-cooling to room temperature (called below ''aged''); and 2-heating to 1100°C with a rate of 10°C per minute, holding at this temperature for 15 min, furnace cooling to 610 8C, holding at this temperature for 8 h followed by aircooling to room temperature (called below ''annealed and aged''). Together with the as-received condition, this provided three conditions to investigate. Heat treatment was performed using an ''Across International'' KTL1400 Tube Furnace. The samples were treated in vacuum.

Materials and experimental techniques
Sample preparation for optical and scanning electron microscopy included mounting in Polyfast resin, polishing on Struers Tegramin-25 automatic polisher to 1 lm finish and etching with ferric chloride solution. Optical microscopy was carried out by using Leica DM 6000 M optical microscope equipped with Leica Application Suite (LAS) 4.0.0 image processing software. More than 400 grains were measured for each material condition to determine average grain size and plot the grain size distributions. Scanning electron microscopy was conducted using JEOL7001F FEG scanning electron microscope (SEM) operating at 15 kV. The energy-dispersive X-ray spectroscopy (EDS) of precipitates was carried out using an AZtec 2.0 Oxford SEM EDS system. Particle compositions were analysed on up to 60 particles for each condition. More than 200 particles were manually measured for each condition to determine their size variation with heat treatment schedule.
Transmission electron microscopy was conducted using JEOL JEM-2011 operating at 200 kV. For the determination of \ 20 nm particles size and number density values, up to 1200 particles have been measured for each studied condition. Particle types were analysed using selected area diffraction technique. The dislocation imaging was conducted near the [101] grain zone axis.
Microhardness was measured on Struers DuraScan Vickers hardness tester with 0.5 kg load. Ten indentations were performed for each condition at a distance of 5-7 mm below the rod surface. The indentation were performed at a distance of approximately five times the length of the indent diagonals to ensure that the results were not contaminated by work hardening from previous indentations. The indentation dwell time was 14 s according to the standard ASTM E384.
Compression testing was carried out on a Kammrath and Weiss GmbH mini-tensile stage (Kammrath and Weiss GmbH, Dortmund, Germany). The testing was performed using cylindrical specimens of 3 mm diameter and 4 mm height cut on a wire cutting machine. Industrial grade grease was applied on top and bottom surfaces of the cylinders to minimise friction between the specimens and the machine traverses, this had to guarantee unidirectional deformation during testing. The constant crosshead speed of 4 lmÁs -1 was applied and resulted in 1 9 10 -3 s -1 strain rate. Two specimens were tested per condition.

Hardness and strength
Hardness and compression yield stress (YS) were measured to be: • 191 ± 8 HV and 320 ± 10 MPa in the as-received hot rolled condition; • 285 ± 5 HV and 470 ± 15 MPa in the aged condition; • 276 ± 7 HV and 360 ± 12 MPa after annealing followed by aging.
As expected, strength increased as a result of heat treatment for both conditions. However, after aging without annealing this increase was higher (hardness grew by 49% and YS by 47%, compared to the asreceived condition) than after annealing followed by aging (hardness grew by 45% and YS by only 13%, compared to the as-received). It is also worth to note that the effect of annealing on hardness was less prominent than this on the yield stress. This difference in hardness and yield stress response to heat treatment could be related to the effect of work hardening during hardness testing. A stronger alloy, the one that was aged without annealing, had a lower capacity for further work hardening during hardness testing (cold deformation), due to a higher dislocation density and particle number density which increase the probability for the dislocation pile-up, formation of tangles, and complete or partial dislocation immobilisation. In contrast, a lower strength alloy, the one that was aged after annealing, exhibited a lower initial (prior testing) dislocation density and a larger capacity for new dislocation generation, this would result in higher work hardening during hardness testing and an increased hardness value. The yield stress value, being the point of variation of the stress-strain curve from a straight line, is not affected by the mechanism of work hardening operating on later stages of tensile testing. To support this explanation, the following microstructure characterisation was carried out.

Grain structure
Due to full solubility of Cu in Ni, microstructure in the studied alloy is single phase (Figure 1). Some deformation and annealing twins have been observed, which is common for Ni alloys that exhibit fcc crystal structure. The grain size, which was measured as a distance between opposite high angle grain boundaries or a grain boundary and a twin in those grain where twins were present, was in the range of 12-80 lm (average 34 ± 16 lm), 12-100 lm (average 36 ± 18 lm), and 30-180 lm (average 77 ± 35 lm) in the as-received, aged, and annealed and aged conditions, respectively ( Figure 2, Table 1). The fraction of [ 60 lm grains significantly increased after annealing and aging ( Figure 2); thus, the average grain size increased by 2.3 times, compared to the as-received condition. This is in line with previously observed grain growth behaviour in Ni-base alloys [31][32][33].

Coarse particles
Optical and SEM imaging revealed precipitation of coarse particles in all studied conditions ( Figure 3). Their sizes ranged from 0.9 to 6.5 lm, from 0.5 to 3.0 lm, and from 0.4 to 1.3 lm in the as-received, aged, and aged and annealed conditions, respectively. The average size of coarse particles slightly decreased from the as-received to aged and to annealed and aged conditions (Table 1). Their number density and  area fraction did not vary significantly between the as-received and aged conditions. However, after annealing and aging the coarse particle number density increased by more than 2 times, compared to two other conditions. This may indicate growth of relatively coarse particles during annealing to sizes visible in SEM, in addition to dissolution during annealing followed by re-precipitation during aging of relatively fine particles. SEM-EDS showed coarse particles to contain Ti, C and N for all material conditions (Figures 4 and 5). Precipitation of TiCN in Monel K500 was observed previously [34,35].

Dislocation density
TEM imaging of dislocation structure exhibited numerous dislocation tangles in the as-received condition (Fig. 6a). Following high temperature processing, the dislocation density decreased (Figure 6b and c, Table 1). Compared to the as-received condition (1.2-1.3 9 10 14 m -2 ), the dislocation density decrease after annealing and aging was by almost 3 times (to 0.4-0.5 9 10 14 m -2 ) and this after aging without annealing was by only 25% (to 0.9 -1.0 9 10 14 m -2 ). This can be explained by high annealing temperature (1100°C) and longer total heat treatment time for the schedule with annealing leading to more developed dislocation annihilation. The dislocation density values observed here correspond to previously measured data for hot deformed and annealed Ni alloys [36,37].

Fine particles
Fine precipitates were observed to be in the size range of 3-11 nm in all the three studied conditions (Figure 6d-f). However, their number density and size distributions varied with processing. Aging resulted in the fine particle growth and precipitation of new particles: (i) the fraction of 3-4 nm particles decreased while this of [ 4 nm particles increased (Fig. 7), which resulted in a slight increase in the average particle size from 3 to 4 nm; and (ii) the total particle number density increased by 23% compared to the as-received condition (Table 1). Following annealing and aging, the fraction of 3-4 nm particles increased and this of [ 4 nm particles decreased (Figure 7), and the total particle number density slightly decreased (by 3%), compared to the as-received condition. Probably, during annealing the fine  A more detailed characterisation of these particles in Monel K500 was presented in our previous work [38]. Diffraction patterns originating from c 0 -Ni 3 (AlTi) particles were observed for the aged and annealed  and aged conditions, but not for the as-received. Possible absence of c 0 -Ni 3 (AlTi) precipitation in the hot rolled alloy could be related to fast cooling rates at the end of thermo-mechanical processing cycle, leaving limited time for nucleation of c 0 particles. The number density of fine particles showed its maximum for the aged condition and minimum for the annealed and aged condition ( Table 1). The variation in c 0 precipitation during ageing, with respect to the prior-to-ageing microstructure, might be related to three aspects: (i) spatial distribution of Al atoms (they could be at a longer distance from each other after annealing), (ii) dislocation density (measured to be lower after annealing), and (iii) texture. The measurements of the unit cell size showed the highest values (0.367 nm) for the annealed and aged condition (Table 1). This may indicate the highest concentrations of solute atoms dissolved in the matrix (note the lowest volume fraction of fine precipitates at this condition). For the Al atoms distributed at a longer distance after annealing, longer times would be required for diffusion towards each other and the particle nucleation. Phase field modelling conducted recently indicated a faster Ni 3 Al nucleation when the bulk Al content was increased [39]. In addition, the density of dislocations is lower after annealing. Thus, Figure 6 Bright field TEM images of dislocation structure and nano-precipitates with corresponding diffraction patterns for the a, d, g asreceived, b, e, h aged, and c, f, i annealed and aged conditions, respectively. the effect of dislocations on pipe diffusion would be lower [40]. The rolling texture is going to change during annealing, and the texture was shown to affect the Ni 3 Al nucleation rate [41], this requires further investigation.

Effect of heat treatment on precipitation and strength
Aging of Monel K500 at 610°C for 8 h resulted in strength increase (Table 1). This was expected due to numerous previous studied of Ni-base alloys showing precipitation strengthening after aging. In particular, coherent Ni 3 Al precipitation was shown for aging temperatures below 700°C [42], and this provides precipitation strengthening and work hardening rate increase [43]. However, in our work the strength increase relative to the hot rolled condition was higher after direct aging compared to aging after annealing. This can be related to annealing leading to grain growth (this reduced the grain boundary strengthening) and dislocation annihilation (this reduced strengthening from dislocation-dislocation interactions). Measurements of coarse precipitates showed close values of area (volume) fraction for the three studied conditions, although due to twofold increase in number density some increase in strengthening contribution from coarse particles might be expected after annealing and aging. Strengthening after aging was expected due to precipitation of TiC and Ni 3 Al particles for both initial conditions. Thus, the measured number density of fine precipitates increased after aging of the hot rolled samples from 0.00025 to 0.00062, and this matched the compression yield stress and hardness increase from 320 to 460 MPa and from 191 to 285HV, respectively. However, for the annealed and aged condition the fine particle number density was lower (0.00019) than in the hot rolled (0.00025), but the yield stress and hardness did increase from 320 to 360 MPa and from 191 to 276HV, respectively (Table 1). With respect to this trend, it is worth to note: (i) an increase in measured unit cell size of the Ni matrix from 0.354 nm for the hot rolled to 0.363 nm for the aged and to 0.367 nm for the annealed and aged conditions, this means enrichment of solid solution with atoms of microalloying elements; (ii) an increase in the number density of coarse particles for the annealed and aged condition, which could provide some precipitation strengthening; (iii) a potential chemistry variation of fine particles for various conditions, as during annealing followed by aging the dissolution and reprecipitation of TiC and Ni 3 Al take place [44,45].

Strengthening mechanisms
To discover strengthening mechanisms operating in Monel K500 and determine their contributions to mechanical properties, the following analysis was carried out.
The grain size strengthening was calculated using the Hall-Petch relationship: where the constants were selected to be r 0 = 21.8 MPa, k = 158 MPaÁ ffiffiffiffiffiffi ffi lm p on the basis of [46][47][48][49].
For the solute atom strengthening, the following equation was used: .] according to [50], and C i is the atomic fraction of each contributing element in solution.
Determination of atomic fractions of elements in solution is a challenge. In this work, we assumed that the total concentrations of Cu, Fe, Mn, and Si in alloy composition, elements that are not expected to form precipitates, were fully dissolved in the Ni matrix. Concentrations of Ti and Al, that are precipitating as TiC and Ni 3 Al, respectively, were assessed using our own theory proposed in one of previous publications [51]. According to this theory: • The volume fraction of TiC, V fTiC , can be expressed as V fTiC ¼ V TiC V Ni , where V TiC is the total volume of TiC particles, and V Ni is the volume of Ni matrix; • Volume of TiC particles is V TiC ¼ a 3 TiC • Volume of the Ni matrix is V Ni ¼ a 3 Ni ÁN Ni 4 , where a Ni is the unit cell size of Ni and N Ni is the number of Ni atoms in the matrix; the number ''4'' specifies the fact that each unit cell of Ni contains 4 Ni atoms (one corner and three faces of the fcc crystal lattice); • Taking into account that N Ti / N Ni is the atomic fraction of Ti atoms that went for formation of TiC particles, substitution of the expressions for TiC and Ni volumes will lead to TiC a 3 Ni .
Solving this equation for the atomic fraction of Ti in TiC, we obtain where V fTiC is a measured volume fraction of TiC particles and a TiC and a Ni are either measured or standard values of the unite cell sizes of TiC and Ni matrix, respectively. The unit cell of Ni 3 Al contains one Al atom in the cube corner and three Ni atoms on the cube faces. Therefore, the volume fraction of Ni 3 Al can be expressed as follows: Ni , and the atomic fraction of Al in Ni 3 Al as To calculate the atomic fractions of Ti and Al in solution, fractions of these elements in the particles should be subtracted from their bulk contents in alloy composition.
For the as-received condition, the atomic fraction of Ti in TiC was calculated as follows: As the Ti concentration in particles (0.844 at%) is equivalent to this in the bulk composition, probably, no Ti was left in solution. On the other hand, Al precipitation requires long aging times, and no Al particle diffraction was observed by TEM. Therefore, all Al was assumed to remain in solution for the asreceived condition.
Aging of the as-received alloy at 610°C was supposed to stimulate precipitation of Ni 3 Al and not affect the TiC particle populations (due to relatively low aging temperature for TiC). The volume fraction of fine Ni 3 Al after aging can be calculated via subtraction of 0.00025 (volume fraction of TiC measured using TEM in the as-received condition) from 0.00062 (total measured using TEM for the aged condition), this will give 0.00037. Substitution of this value gives Al fraction in Ni 3 Al particles. During aging after annealing both particle types, TiC and Ni 3 Al precipitated from solution with 0.0151 total area fraction (for large numbers is equal to volume fraction) of the coarse and 0.00019 total volume fraction of fine particles. Assuming full precipitation of Ti in the form of TiC, the TiC atomic fraction can be calculated as follows: Ti atfr ¼ f TiC The absence of free Ti after precipitation of coarse TiC allows to assign the whole volume fraction of fine particles 0.00019 to represent Ni 3 Al.
Thus, the atomic fraction of Al in the particles will be. Dislocation looping around coarse incoherent precipitates may provide the precipitation strengthening contribution calculated using the Ashby-Orowan equation [52]: where, G, b, f and X are shear modulus, Burgers vector, particle volume fraction, and average particle diameter in micron.
After substituting G = 79,000 MPa and b = 0.352 nm (for pure Ni), Eq. (5) will look for Ni in the following form: Eq. (6) was used to calculate precipitation strengthening from TiC particles in both SEM ([ 20 nm) and TEM (\ 20 nm) size ranges. The dislocation looping around relatively large particles ([ 20 nm size) occurs due to the lower energies required for looping compared to cutting. The dislocation looping around relatively small (\ 20 nm) TiC was assumed in this work on the basis of high lattice mismatch (15-18%) between TiC (0.433 nm of unit cell size) and Ni (0.354-0.367 nm measured in this work) that would make cutting less favourable.
Dislocation cutting of fine coherent precipitates may provide the precipitation strengthening contribution calculated using three major approaches [53]: -(Particle atomic) order strengthening where M = 3 is the matrix orientation factor, b = 0.354-0.367 nm is Burgers vector for the Ni matrix, c is the matrix-particle interface energy assumed for the Ni-Ni 3 Al interface to be c = 0.04 Jm -2 [54,55], and f is the particle volume fraction; -Lattice mismatch (between the particles and matrix) Where a = 1 is a constant, r is the average precipitate radius, G = 79,000 MPa is the shear modulus of Ni at room temperature, e ¼ a Ni3Al Àa Ni a Ni is the constrained lattice parameter mismatch, a Ni-= 0.354…0.367 nm and a Ni3Al = 0.357 nm [54][55][56]; -Share modulus mismatch (between the particles and matrix) Where DG = G Ni3Al -G, G Ni3Al = 207,000 MPa, r is the average particle radius.
The work hardening contribution from dislocation-dislocation interactions to the yield stress was estimated using the long range work hardening theory [57]: Where a = 0.5 is a constant, and q is the dislocation density.
The total yield stress would be the summary of contributions from microstructural parameters: Substitution of all the measured microstructural parameters, namely: average grain size to Eq. (1); Cu, Fe Mn, Si, and Al concentrations in solution to (2), volume fraction and average diameter of coarse particles to (6), volume fraction and average radius of fine particles to (7)(8)(9), and the dislocation density to (10), gave the microstructural contributions to the yield stress for the three studied conditions (Table 2, Figure 8).
Analysis of Table 2 and Figure 8 brought the following results: • the error between measured and calculated values of yield stress was in the range of 6-10%, which indicates a reasonable accuracy of equations and assumptions considered in the calculation; of course, a certain level of local inhomogeneity would affect the measured average values of grain size, particles parameters, and dislocation density; theories applied in this work for calculation of precipitation strengthening contributions did not take into account the effect of particle shape, although the particle shape does have an effect, especially on the interaction of moving dislocations with fine coherent particles [58]; all the dislocations measured in this work were considered equally contributing to strengthening; however, their type and spatial configuration influence the flow behaviour of metallic materials [59]; • the compression yield stress in the hot rolled and aged material was by 150 MPa (47%) higher than in the as-received, and this in the annealed and aged was by only 40 MPa (12%) higher than in the as-received; this trend matched for both measured and calculated values; • In the hot rolled as-received alloy, the precipitation strengthening contribution from fine TiC particles was significant (40%), although the grain size (17%) and solid solution (31%) strengthening effects provided substantial contribution; • In the hot rolled and aged alloy, the prevailing contribution was coming from precipitation of fine TiC and Ni 3 Al particles (66%); • In the annealed and aged material, the strengthening contribution from precipitation of fine particles was lower (53%) compared to the hot rolled and aged, and the solid solution contribution (26%) was reasonable; • The solid solution strengthening was high (88 MPa) for all the studied conditions, although loses in effectiveness to precipitation strengthening; • Amongst the six elements considered here (namely Fe, Mn, Si, Cu, Al, and Ti), Ti shows the highest strengthening coefficient (775 MPa / ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi at: fraction p ]) but seems to be more affective in TiC precipitation; this corresponds to recently discovered potential of Ti-rich particle precipitation to increase strength of (TiN ? TiB)Ti composites by 85% [60];  Figure 8 Relative contributions to the yield stress from the microstructural parameters for the three studied alloy conditions.
• Solute Mn, Si and Fe are of medium strengthening potential; and Al is definitely more effective in Ni 3 Al precipitation rather than in solution; significant strengthening from Al-and Ni-rich particles was recently reported for Cu-base and high entropy alloys produced by various methods of additive manufacturing [61,62]; • The effect of dislocation-dislocation interactions was minor in the studied alloy and decreased after aging, although it is worth to note a twice larger dislocation strengthening contribution (8%) for the as-received hot rolled condition compared to the aged (4%).
Considering the effects of processing technology on strengthening mechanisms operating in Monel K500, several points can be discussed: • Grain coarsening, that may occur during high temperature annealing, decreases the contribution from grain size strengthening; however, an increase in annealing twin density, which usually occurs in fcc metals during long-term holding at elevated temperatures [63][64][65], will reduce the dislocation free path in those grains were twins form; therefore, the effect of twins on strengthening should be taken into account [66]; • A higher dislocation density in the hot rolled alloy, compared to the annealed, facilitates pipe diffusion of alloying elements and particle nucleation during aging, thus the number density of fine TiC and Ni 3 Al in the hot rolled and aged alloy was higher than in the annealed and aged; a higher number density of finer particle produced significant precipitation strengthening effect; • Calculations showed that only 0.01 at% of Al out of 7.07 at% in the alloy composition precipitated in the form of Ni 3 Al in the hot rolled and aged condition, this provided a quite high 46% contribution to strengthening; probably, aging parameters (temperature and time) could be optimised to obtain a larger density of Ni 3 Al and increase the precipitation strengthening effect utilising the same Al content; • Populations of fine TiC particles produced during hot rolling may be lost during annealing, and their contribution in the hot rolled and aged condition (20%) was noticeable; • It is worth to minimise the loss of Ti and Al to coarse precipitates, as their number density and potential interactions with dislocations are low; this can be achieved via optimisation of the hot rolling schedule (which is outside of the paper scope).

Conclusions
Investigation of the Monel K500 (NiCuAlTi alloy) microstructure and strength variation with heat treatment brought the following conclusions: 1. Following aging at 610°C for 8 h, the hardness and compression yield stress of the initially hot rolled alloy increased by 94 HV (49%) and 150 MPa (47%), respectively; and these of the annealed alloy increased by lower values of 85 HV (44%) and 40 MPa (12.5%), respectively. This discrepancy is related to the negative effects of annealing on grain size strengthening (due to grain growth), work hardening (due to dislocation annihilation), and precipitation strengthening (due to particle dissolution). 2. In the hot rolled alloy, the precipitation strengthening contribution from fine \ 20 nm TiC particles was significant (115 MPa or 40% of total), although the grain size (49 MPa or 17%) and solid solution (88 MPa or 31%) strengthening effects also provided valuable contributions. Aging of the hot rolled alloy increased the precipitation strengthening contribution by almost 3 times (to 331 MPa or 66%) via precipitation of Ni 3 Al particles in addition to TiC. 3. In the annealed and aged alloy, the strengthening contribution from precipitation was by 46% lower (178 MPa or 53%) compared to the hot rolled and aged condition, which is in line with a lower number density of fine particles in the annealed and aged condition. This could be related to (i) dissolution of \ 20 nm TiC during annealing, and (ii) retardation of \ 20 nm Ni 3 Al precipitation during aging. The latest would be due to a lower dislocation density in the annealed alloy compared to the hot rolled, that slows down the pipe diffusion. 4. Our calculations showed fine Ni 3 Al to be more effective strengthening agents than TiC particles of similar size and number density. This could be explained by the variation in dislocation-particle interaction mechanisms: dislocation looping around TiC requires less energy than dislocation cutting through Ni 3 Al, thus leading to lower strengthening from TiC compared to Ni 3 Al. 5. Our calculations showed a very minor amount (0.01 at%) of total Al content (7.07 at%) precipitating during aging of hot rolled alloy at 610 C for 8 h. Nevertheless, the precipitation strengthening contribution from only Ni 3 Al was significant (228 MPa). Modification of the heat treatment schedule could further increase the precipitation strengthening effect of Al alloying. 6. Amongst the six elements considered in this work to provide the solute atom strengthening, Ti showed the highest strengthening coefficient but was more affective in TiC precipitation; Mn, Si and Fe exhibited medium strengthening potential; and Al was more effective in Ni 3 Al precipitation rather than in solution.

Acknowledgements
This paper includes research that was supported by DMTC Limited (Australia). The authors have prepared this paper in accordance with the intellectual property rights granted to partners from the original DMTC project. The experiments were carried out at the University of Wollongong. The JEOL JEM-2011 TEM and JEOL JSM-7001F FEG-SEM used in this work for microstructure characterisation were funded by the Australian Research Council grants LE0237478 and LE0882613.

Funding
Open Access funding enabled and organized by CAUL and its Member Institutions.

Data availability statement
The data presented in this study are available on request from the corresponding author.

Declarations
Conflict of interest The authors declare that they have no competing interests.
Ethical approval No ethical approval was required for this research as it did not involve human tissue or any other parts of living organisms.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/.