100 Particles Quantum Heat Engine: Exploring the Impact of Criticality on Efficiency Anass Hminat, Abdallah Slaoui, Brahim Amghar, Rachid Ahl Laamara Abstract Quantum many-body systems present substantial technical challenges from both analytical and numerical perspectives. Despite these difficulties, some progress has been made, including studies of interacting atomic gases and interacting quantum spins. Furthermore, the potential for criticality to enhance engine performance has been demonstrated, suggesting a promising direction for future investigation. Here, we explore the performance of a quantum Otto cycle using a long-range Ising chain as the working substance. We consider an idealized cycle consisting of two adiabatic transformations and two perfect thermalizations, eliminating dissipation. Analyzing both engine and refrigerator modes, we investigate the influence of particle number, varied from 10 10 to 100 100 , on efficiencies and behavior near the critical point of the phase transition, which we characterize using a scaling factor. We also examine how internal factors—specifically, the power-law exponent, the number of particles, and the hot and cold reservoir temperatures—affect the system’s operation in different modes. Our results reveal that these factors have a different impact compared to their classical counterparts. I Introduction Thermodynamics is a remarkable theory in physics, primarily aimed at studying heat Winterbone2015 . It has been instrumental in the design of internal combustion engines and even in explaining the behavior of black holes. Technological progress increasingly requires the miniaturization of components Deffner2019 , pushing them to a regime where quantum effects are no longer negligible. The necessity of understanding the underlying mechanisms of converting heat into useful work has led researchers to formulate the three laws of thermodynamics. More recently, with the rise of quantum technologies and the miniaturization of devices that exchange heat and work at the nanoscale, it has become increasingly relevant to investigate these mechanisms within the framework of quantum mechanics Strasberg . However, while classical thermodynamics is a well-established theory, its extension to the quantum domain presents conceptual challenges and remains an active area of research. Pioneering efforts to generalize the concepts of heat and work for quantum systems date back to the 1980s Perez . Only recently has a renewed interest in quantum thermal machines spurred prolific scientific advancements, starting with the groundbreaking proposal of the maser Scovil1959 as the first example of a quantum machine. A quantum thermal machine is broadly defined as a device made of quantum systems capable of performing work by undergoing a suitable thermodynamic cycle. Several studies have proposed potential implementations of few-body thermal machines Latifah2013 based on the Carnot Bender2000 ; Bender2002 , Otto Solfanelli2020 ; Piccitto2022 , Stirling Yin2017 ; Yin2018 , Brayton Setyo2018a , and Diesel cycles Setyo2018 . Some of these realizations have been achieved using various quantum systems, including infinite potential wells Purwanto2016 , nuclear magnetic resonance (NMR) Camati , ultra-cold atoms Bouton2021 . However, the role played by many-body interactions in the thermodynamic performance of quantum machines Arezzo2024 ; MakouriSlaoui2025 ; MakouriSlaoui2024 is still not fully understood. Determining whether quantum machines with many-body interactions can outperform classical ones, and under what conditions this occurs, is currently at the forefront of debates in quantum thermodynamics. Long-range interacting systems Ranabhat2023 ; Amghar2023 offer promising prospects for quantum technological applications due to their resilience to external disturbances Lerose2019 . This stability enables control over the impact of dynamically generated excitation, thereby mitigating their detrimental effects. Specifically, the dynamical phase transition Piccitto2019 ; Piccitto2019a is clear when \alpha<1 . But when \alpha>1 , the critical point grows into a chaotic crossover region where the dynamics and asymptotic state are very dependent on the system’s parameters. A promising direction for exploring quantum advantage lies in using many-body quantum systems as the working substance in thermodynamic engines MurherjeeV ; CangemiLM . Although this approach is hindered by technical challenges, both analytically and numerically, some preliminary results have been obtained JamilloM ; WatsonRS ; Halpernny . It has been shown that quantum criticality can play a crucial role in enhancing engine performance Carollof ; Zhang2007 ; Munoz2014 . However, the exact role that many-body interactions play in the thermodynamic efficiency of quantum engines remains poorly understood. Besides, determining whether quantum engines featuring such interactions can outperform classical ones, and under what conditions, is a key question in the current quantum thermodynamics debate. One of the main obstacles lies in the difficulty of solving the dynamics of these many-body systems. In this paper, we contribute to this burgeoning field by presenting an in-depth study of the quantum Otto engine based on the Ising model (see Fig. 1 ), focusing particularly on the potential influence of quantum criticality on the engine’s thermodynamic performance. This article is structured as follows: First, we will present our system and the Otto cycle, describing its applications. We will explore the impact of internal parameters such as N , h_{i} , \alpha , \beta_{c} , and \beta_{h} on the operating modes and highlight the quantum features absent in classical systems, as well as the transition between them. Next, we will look at the quantum refrigerator (or thermal engine) mode and see how the chain size (N) and thermal correlations change the quantum refrigerator capability Q_{c}/N (or W/N ) and the efficiency \eta_{R} (or \eta ). We will fix the time of adiabatic compression throughout the study, unlike in previous works. We will address the scenario of a perfect quantum thermodynamic machine and a refrigerator, both of which experience an infinitely slow thermodynamic phase. We will calculate the work per spin, the efficiency, and the influence of N on the extraction of work per spin and its effect on efficiency. We will also perform the same analysis for the refrigerator and also describe the performance and the influence of perturbations and dissipation, using the scaling factor Campisi2016 \Pi_{R}/N (or \Pi/N ), which serves as a reliable indicator for stability. In the following section, we will examine the behavior of correlations Piccitto2019 ; Piccitto2019a for \alpha ranging from 0.01 to 1.3 with a fixed N=100 , focusing on quantum refrigerators and thermal engines. We will study how the efficiency behaves as a function of Q_{c}/N and h_{i} (or \eta as a function of W/N and h_{i} ), and analyze their distribution across chains of varying sizes. Our final step is to calculate the proportionality between the scaling factor per spin for N from 10 to 100 , as well as the behavior of peaks before and after the phase transition and at the critical field. II quantum otto cycle In this section, we will describe the system and outline its phase transition. We will then examine the thermodynamic cycle and the associated physical quantities, such as the heat exchanged and the work done. Afterward, we will explore how the interaction range affects the efficiency of the cycle. The analysis will focus on how both short-range and long-range interactions influence the system’s performance, particularly near the critical point. II.1 Many-body quantum Ising model We study the one-dimensional transverse-field Ising model with long-range (LRIM) power-law decaying interactions. We consider an open boundary condition, which is described by the following Hamiltonian H=-\frac{1}{K(\alpha)}\sum_{i0 , in the presence of magnetic field h . Hereafter, we set J=0,5 as the energy scale and work in units where \hbar=k_{B}=1 . Figure 1 : The schematic of the Ising Long-range quantum Otto engine consists of two adiabatic phases and two thermalization phases. From (A → B), the system undergoes an adiabatic process where the transverse field increases from h_{i} to h_{f} . In (B → C), the system thermalizes at temperature T_{h} . Then, from (C → D), it is adiabatically brought back to h_{i} , followed by thermalization at temperature T_{c} in (D → A), completing the cycle. The long-range interaction between spins is dictated by the strength parameter \alpha (we set h_{i}70 ), the work per spin is higher in the region where h_{i}\leq 1.0 . Beyond this point, the slope decreases, leading to a stabilization of the curve. The maximum work per spin, W/N , occurs around h_{i}=0.25 , indicating a more effective capture of short-range correlations near the critical point. It is noteworthy that, across all regions, long chains exhibit a higher W/N compared to short chains. After the critical point, W/N declines rapidly, especially for large systems, signaling a significant shift associated with the phase transition. In contrast, for smaller systems, the maximum is less pronounced, and the work is lower, highlighting the dominance of thermal fluctuations and local correlations. When comparing the sizes of the chains, W/N remains higher for long chains throughout the h_{i} range from 0 to 2.5, demonstrating greater efficiency in capturing short-range correlations for longer chains. In Fig. 10 ( b ), corresponding to \alpha=1.2 , \beta_{h}=0.2 , and \beta_{c}=0.9 , the work per spin is similarly elevated for larger values of N , peaking around h_{i}=0.35 , which aligns with the estimated critical point for \alpha>1 at h=0.7 . Long chains are more adept at capturing short-range correlations than their shorter counterparts prior to the transition, thereby accounting for a higher W/N . After h_{i}=0.7 , W/N experiences a steep decline, occurring more rapidly than in the previous case, while the work per spin of long chains remains superior to that of short chains. This observation indicates that longer chains continue to effectively capture short-range correlations even subsequent to the phase transition. In contrast, shorter chains exhibit a more stable behavior, suggesting reduced sensitivity to fluctuations in h_{i} . The Fig. 10 ( c ), with \alpha=0.55 , \beta_{h}=0.2 , and \beta_{c}=0.667 , shorter chains ( N<30 ) exhibit a higher W/N for h_{i} ranging from 0 up to the critical point at h=1.0 , where a transition occurs, allowing long chains to become more efficient following the phase transition; this alteration does not apply to short-range correlations. The peak of W/N is observed around h_{i}=0.25 , indicating that smaller chains are more effective in capturing intermediate-range correlations prior to the phase transition. However, after h_{i}=1 , longer chains ( N>70 ) surpass shorter chains in terms of W/N , reflecting a superior ability to capture long-range correlations. This implies that for \alpha=0.55 , the phase transition significantly alters the behavior of long chains post h_{i}=1 , while shorter chains maintain their efficiency before this transition. Fig. 10 ( d ), with \alpha=0.2 , \beta_{h}=0.6 , and \beta_{c}=2.0 , exhibits behavior akin to the previous case, with the performance of long chains surpassing that of shorter chains before the critical point at h=1 , particularly near h=0.4 . For shorter chains ( N<30 ), W/N is elevated for h_{i}\leq 0.4 , peaking around h_{i}=0.8 for longer chains. Beyond this threshold, W/N declines rapidly, indicating a loss of efficiency in the smaller chains. In contrast, after the critical field, starting from h_{i}=1 , there is a significant drop in W/N . This illustrates that long-range correlations in long chains at the phase transition accelerate the decline of W/N . The contour lines of the work per spin W/N presented in Fig. 11 show behavior that depends on \alpha and the thermal gap. When \alpha is small and favors long-range correlations, there are two peaks: one for short chains and the other for long chains unlike \alpha=0.25 , we observe a peak before the transition (for small N ) and another at the critical field (for large N ). When \alpha=0.55 , there is a peak before the critical field and another after the phase transition, which is farther apart with a lower W/N at the critical field compared to the previous case. When short-range correlations are favored ( \alpha=0.9 and 1.2 ), we observe a single peak before the critical field for long chains. The only difference between the two cases lies in the variability and a smaller decrease in W/N before the phase transition for \alpha=1.2 , with more vertical contour lines. This detailed analysis presents the behavior of the efficiency W/Q_{h} and the work per spin W/N as a function of h_{i} for different values of \alpha , \beta_{h} , and \beta_{c} . Figure 11 : Contour plot of Work per spin W/N and efficient \eta as a function of h_{i} and N. With \alpha=0.9 , \beta_{h}=0.5 , and \beta_{c}=1.0 , the efficiency W/Q_{h} shows a gradual decline as h_{i} increases. The phase transition, occurring at h=1 , corresponds to the critical point h_{c} for \alpha\leq 1 . Prior to this transition, longer chains (with lower N ) display greater efficiency compared to shorter chains. However, near the critical point, the efficiency of longer chains aligns closely with that of shorter ones. Beyond h=1.25 , the disparity between short and long chains becomes more apparent, accompanied by a marked decrease in efficiency for both. In the second figure, with \alpha=1.2 , \beta_{h}=0.2 , and \beta_{c}=0.9 , we observe a pattern reminiscent of the case with \alpha=0.9 , where short-range correlations are predominant. The critical field is located at h_{i}=0.7 . Prior to the phase transition, there is a slight reduction in efficiency for chains of all sizes, with large N chains exhibiting greater efficiency. At the critical point, a distinctive behavior emerges with a sharp drop in efficiency, resulting in a significant decline across all chains, leading to an identical efficiency independent of chain length in the post-critical region. Efficiency remains relatively high for small values of h_{i} (between 0 and 0.5), reaching a peak at very low fields. Large chains (with higher N ) demonstrate superior efficiency compared to smaller chains across the entire range of h_{i} , both before and after the critical field. This stability in larger chains following the transition highlights a regime where short-range correlations prevail, maintaining efficiency despite variations in the transverse field h_{i} . In the third figure, with \alpha=0.55 , \beta_{h}=0.2 , and \beta_{c}=0.667 , which favors intermediate range correlations, the efficiency W/Q_{h} exhibits a gradual decline as h_{i} increases, with nearly identical efficiency for large N values, indicating pre-saturation for long chains in the range h_{i}=0 to 0.5, just before the phase transition. Around the critical field h_{c}=1 for \alpha\leq 1 , a stabilization is noticeable, where efficiency remains almost unchanged. Beyond h_{i}=1.5 , a slight reduction in efficiency is observed for all chain lengths. The curve’s behavior suggests heightened sensitivity of the system to changes in h_{i} near the critical point. The Fig. 12 ( a ), with \alpha=0.2 , \beta_{h}=0.6 , and \beta_{c}=2.0 , the efficiency is initially high but steadily declines as h_{i} increases, reaching the critical point at h_{i}=1 . This behavior reflects a phase transition, where the system’s sensitivity to the transverse field h_{i} is notably heightened for small values of \alpha . The slope increases and the decline becomes more rapid for larger h_{i} , emphasizing the influence of long-range correlations in stronger fields. Post-transition, shorter chains demonstrate greater stability with fewer fluctuations compared to longer chains. For large N , the efficiency becomes more uniform, exhibiting faster saturation, more pronounced than in the case of \alpha=0.55 , though this saturation occurs in the post-transition region, accentuated by long-range correlation effects. In stronger fields, such as h_{i}=2 , intermediate-length chains show superior efficiency, indicating a shift in the system’s operational regime where long-range correlations become less effective for longer chains (see Fig. 12 ). Figure 12 : Efficiency \eta as a function of h_{i} for different N . The contour lines of efficiency \eta show a dependence on \alpha . Similar to W/N , there are two distinct behaviors. For small \alpha , we observe a peak for long chains at very low h_{i} , with the contour lines becoming more horizontal for medium and large N , then more vertical for long chains, which is less noticeable for \alpha=0.55 . For larger \alpha , there is a clear peak for medium and large chains at low h_{i} , with vertical contour lines where efficiency increases horizontally. This effect is more pronounced for \alpha=1.2 , which shows slower variability compared to \alpha=0.9 (Fig. 11 (e)-(h)). Unlike work, the efficiency for all the parameters previously mentioned is dependent on N . To increase efficiency for fixed parameters, it is sufficient to increase the number of particles, which is not always the case for W/N . We observe that the efficiency ( \eta ) and the work per spin ( W/N ) exhibit different behaviors, as the work per spin is not proportional to the chain size for small and intermediate values of \alpha . In contrast, the efficiency depends directly on the size of the chain: the larger N becomes, the greater the efficiency ( \eta ). This discrepancy can be explained by the nature of the correlations and fluctuations induced by the temperature difference between the reservoirs. For larger values of \alpha (e.g., 0.9 and 1.2), we observe that when the temperature difference ( \Delta T ) is significant (as with \alpha=1.2 ), the system achieves a higher work per spin ( W/N ) compared to \alpha=0.9 , but with a lower efficiency. This can be attributed to thermal fluctuations and agitation, which are reduced when \Delta T is smaller. The same interpretation applies to long-range and intermediate-range correlations, where the effect of temperature fluctuations diminishes with a smaller temperature gradient, leading to different performances in terms of both work and efficiency. Thus, the relationship between efficiency and work per spin reflects the complex interplay of thermal fluctuations and correlations in the system, especially across different interaction ranges. III.3 Scaling factor of the quatum heat engine The scaling factor \Pi=\frac{W}{\eta_{\text{Carnot}}-\eta} measures how effectively an engine produces work despite having an efficiency lower than the Carnot limit. A high scaling factor means that the system is optimized to extract significant work, even in non-ideal conditions. In quantum engines, a higher \Pi reflects more efficient energy conversion through quantum effects like correlations and superposition. It also suggests lower energy losses, allowing the system to approach Carnot efficiency. Overall, a high scaling factor indicates that the system is productive, efficient, and resilient to fluctuations. To observe the impact and influence of correlations without considering the size of the chain, we will normalize by dividing by N .We will discussed this in more detail in the appendix Appendix D . The corrected interpretation of the provided figures focuses on the scaling factor \Pi , defined as \Pi/N=\frac{W}{(\eta_{\text{Carnot}}-\frac{W}{Q_{h}})}/N, (8) where W is the work done by the system, \eta_{\text{Carnot}}=1-\frac{\beta_{h}}{\beta_{c}} is the Carnot efficiency, and N is the number of particles in the system . The figures 13 illustrate \Pi/N as a function of the transverse magnetic field h_{i} , highlighting the behavior of the system for varying values of N , ranging from 10 to 100, for different configurations of \alpha , \beta_{h} , and \beta_{c} . Fig. 13 ( a ) \alpha=0.9 , \beta_{h}=0.5 and \beta_{c}=1.0 , \Pi/N starts at high values for small h_{i} and decreases significantly near the critical magnetic field h_{i}=1 , indicating a strong phase transition. Large chains ( N>70 ) show a sharp decrease in \Pi/N , while small chains show a more gradual decline. After h_{i}=1 , \Pi/N for large chains remains stable but lower compared to small h_{i} values, indicating a more consistent capture of long-range effects. \Pi/N drops significantly after the transition, reflecting their inability to maintain high performance in the long-range regime. In , corresponding to \alpha=1.2 , Fig. 13 ( b ) \beta_{h}=0.2 and \beta_{c}=0.9 , \Pi/N is high for large values of N (for example, N>70 ) in the region where h_{i}\leq 0.7 . The maximum occurs around h_{i}=0.3 , reflecting better adaptation to critical fluctuations near the estimated transition point for \alpha>1 . Beyond the critical point h_{i}=0.7 , there is a sharp drop in the system’s performance, with a less steep slope for small chains. After h_{i}=1 , \Pi/N decreases for large chains but remains higher compared to small chains ( N<30 ), indicating that large chains are more effective in maintaining high performance before the transition but experience a more intense drop in performance than small chains, where long-range correlations dominate. Large chains show a rapid decrease in \Pi/N , reflecting their limited capacity to harness critical effects. At \alpha=0.55 Fig. 13 ( c ), \beta_{h}=0.2 and \beta_{c}=0.667 , shows that \Pi/N is initially high, especially for small chains. Where long-range correlations are moderate before the critical field h=1 , in the ferromagnetic phase, they force the particles to align better than in long chains, thus ensuring better \Pi/N performance. After the transition phase h_{i}=1 , large chains dominate in terms of \Pi/N , reflecting their superior performance in exploiting intermediate correlations in the ferromagnetic phase at higher field strength. This behavior indicates that large chains are more adaptable and efficient after the transition, whereas small chains are more efficient in the ferromagnetic phase, before the phase transition. In Fig. 13 ( d ), with \alpha=0.2 , \beta_{h}=0.6 and \beta_{c}=2.0 , where the interactions are the longest, we observe that before h_{i}=0.25 , the short chains perform better, and there is a shift at this point towards the dominance of long chains. \Pi/N peaks at h_{i}=0.6 before the critical point but decreases rapidly beyond this point, showing poor performance in the post-transition regime. We also observe that \Pi/N decreases slowly after h_{i}=1 , indicating an effective capture of long-range correlations. This highlights the dominance of long-range effects for large chains before the phase transition. It is also important to note that, as in the previous section, there is a slight shift between long and short chains post-phase transition, as the short chains stabilize more quickly because the fluctuations and thermal perturbations, amplified by long-range correlations in the paramagnetic phase. Figure 13 : Contour plot of the scaling factor per spin \Pi_{R}/N as a function of h_{i} and N. The impact of \alpha , \beta_{h} , and \beta_{c} on the system is significant. For high \alpha values (e.g., \alpha=1.2 ), short-range correlations dominate, leading to a smoother transition and more stable \Pi/N . Large chains benefit significantly from short-range interactions before h_{c} . For low \alpha before critical field long-range effects are pronounced, leading to sharper transitions in \Pi/N and small chains are more effective in these regimes, showing larger values of \Pi/N . A higher thermal gradient enhances the sensitivity of \Pi/N to variations in h_{i} . For example, in the third figure, the large difference between \beta_{c}=2.0 and \beta_{h}=0.6 causes a significant shift in \Pi/N post-transition.faor a small thermal gap (e.g., \beta_{c}=1.0,\beta_{h}=0.5 in the second figure) result in a smoother decline of \Pi/N , with less abrupt changes across the critical point but with a lower performance than the others. Before the critical point ( h_{i}<1 ), small chains exhibit higher \Pi/N , reflecting their ability to efficiently capture long-range correlation in weak field h_{i} , while large chains have a lower \Pi/N but show a better performance than the shorter chain for higher h_{i} special in the vicinity of critical point and after phase transition, indicating the growing influence of long-range correlations after the critical point ( h_{i}>1 ), large chains outperform small chains, with a higher \Pi/N due to the dominance of long-range correlations. Small chains show a significant decline in \Pi/N , reflecting reduced efficiency in capturing long-range effects. At the critical point ( h_{i}=1 ), \Pi peaks for large chains as the system approaches Carnot efficiency, reflecting enhanced work output and system efficiency. For small chains, the peak is less pronounced, indicating their limited capacity to utilize critical phenomena effectively. ; Figure 14 : The evolution of the maxima as a function of N (the number of particles) for the diamagnetic peak (in red) and the paramagnetic peak (in blue) of \Pi/N and \Pi_{R}/N , with a comparison to the critical field (in red). For the case of short-range correlations, we observe an underperformance after the phase transition, but with a stronger dependence on system size, as longer chains prove more efficient both before and after the critical point. It is also noted that for large \alpha values, short-range correlations dominate, and the system’s efficiency is enhanced by the thermal gap, which promotes short thermal correlations and better protects them from thermal fluctuations. The behavior of \Pi/N as a function of h_{i} demonstrates the intricate relationship between work, system efficiency, and particle number N . Large chains tend to dominate in regimes where long-range correlations are significant, especially beyond the critical point, whereas small chains are more effective in the short-range, pre-transition regime. The parameters \alpha , \beta_{h} , and \beta_{c} play crucial roles in shaping this behavior, with their values determining the system’s sensitivity to field variations and the overall efficiency relative to Carnot. Understanding these dynamics is essential for optimizing quantum heat engines across different operating regimes. IV Investigating the role of thermal correlations and criticality In this section, we will examine the scaling factor \Pi/N , studying its impact on the system’s behavior for varying values of N (ranging from 10 to 100) and different configurations of \alpha , \beta_{h} , and \beta_{c} . focusing on the behavior near the critical field, including the peaks observed in the paramagnetic and ferromagnetic phases. Additionally, we will discuss how these features depend on N in both the quantum heat engine and refrigerator modes. IV.1 Critical reaction We will examine the role of the scaling factor and the behavior at the critical field for N ranging from 10 to 100. We will also analyze the peaks in the paramagnetic phase before the phase transition and the ferromagnetic phase after the transition, as well as their proportionality and dependence on N for the quantum heat engine and refrigerator mode. We intend to use a regression to calculate the exponent a responsible for the growth rate of the scaling factor per spin, proportional to N^{a} . We will then examine the behavior of the performance and its trend for small, intermediate, and long chains and the role of thermal gaps and correlations \alpha , to observe the impact of the thermal reservoirs on our system. In the case where \alpha=0.9 , with \beta_{h}=0.5 and \beta_{c}=1.0 Fig. 14 ( a ), the obtained exponents are a=0.4578 for the critical field, and 0.5154 and 0.46 respectively for the ferromagnetic and paramagnetic phases , with a poor temperature difference between the reservoirs. The scaling factor per particle describes the performance per particle while accounting for the influence of others. We observe that the post-critical and critical exponents are almost identical, with a significant slope for small N (10 to 40) and medium N (40 to 70). This suggests that critical effects slow down the growth and stabilize it in the paramagnetic phase, especially for large N (70 to 100). In the ferromagnetic phase, we see greater growth, with a slight decrease in slope for larger N , and the curve is higher than in the critical and paramagnetic cases. For \alpha=1.2 , with \beta_{h}=0.2 and \beta_{c}=0.9 Fig. 14 ( b ), the obtained exponents are a=0.2355 for the critical field, 0.2308 and 0.2355 for the ferromagnetic and paramagnetic phases , respectively. In this case, \alpha>1 , which favors more short-range correlations, and the thermal gap is higher compared to the previous case. The exponents are smaller than for \alpha=0.9 and are very close to each other, with one of the curves (critical and post-critical) perfectly overlapping The paramagnetic phase exhibits slightly lower growth than the ferromagnetic phase, but both show faster growth with N . A similar behavior is observed for small to intermediate N , but a notable difference appears for large N , where the critical and paramagnetic curves show the onset of pre-saturation the slope increases only marginally between N=90 and N=100 . This indicates that for larger \alpha , the system’s performance depends on its size, which significantly amplifies thermal fluctuations and correlations, potentially leading to a more optimal yield. For Fig. 14 ( c ) \alpha=0.55 , which favors intermediate correlations, with a high thermal gap \beta_{h}=0.2 and \beta_{c}=0.667 . Before the critical field, we observe a curve that is higher than the critical field and the post-phase transition, but it shows a downward behavior indicating underperformance of intermediate chains, which slows down around N=90 . Near N=55 In the ferromagnetic phase, the spins are better aligned, but increasing the number of particles makes the system more sensitive to fluctuations caused by the different types of correlations, which increase the noise and cause the decrease of the scaling factor per spin ( \alpha=-0.1035 ). There is an overestimation of the scaling factor per spin of the post-critical field peaks ( \alpha=0.2432 ), with an increase of the performance of long chains. At the critical point, there is a slight and slow increase for long chains ( \alpha=0.0504 ), indicating a greater difficulty aligning the particles for intermediate correlations before and during the phase transition. In the case of \alpha=0.2 , with \beta_{h}=0.6 and \beta_{c}=2.0 , Fig. 14 ( d ) the thermal gap is short, low values of \alpha favor long-range correlations, indicating that the fields before the critical point are more efficient. For all chains, performance in the phase transition and after the critical field shows a slight decrease in intermediate N, while the curves for the critical and post-critical peaks demonstrate an increase and outperform of the high chains, which is proportional to N .A slight under performance is observed in long chains (from 70 to 100 ) before the phase transition. The scaling factor per spin in the critical fields and after the transition the curves are superimposed for intermediate and shorter chains. This indicates that in long correlation before the critical field there is a superior performance and better spin alignment, resulting in less dissipation and a higher efficiency, there is nonetheless slower growth ( \alpha=0.32826 ) compared to postcritical fields ( \alpha=0.9434 ) and at the phase transition ( \alpha=0.9413 ). We will then move on to the refrigeration mode, which is more prevalent in regions with a small thermal gap. We will examine the influence of the parameter \alpha , considering both small and large values, which, respectively, promote long-range and short-range thermal correlations. In the case of \alpha=0.25 , with \beta_{h}=1.4 and \beta_{c}=1.5 , Fig. 14 ( e )there is a perfect superposition for intermediate and long chains, with a slight difference for small N for the critical ( \alpha=1.6150 ) and postcritical fields ( \alpha=1.2702 ). We observe that the performance per particle of the post phase transition for all N is lower than the pre phase transition, which exhibits a significant increase for small chains and saturates at intermediate N ( \alpha=0.5338 ), then slightly decreases. In contrast, the critical field reduces the gap for larger chains. It is also noted that the performance before the phase transition is less dependent on the size, which is an effect of long-range correlations. For of \alpha=1.2 , with \beta_{h}=0.58 and \beta_{c}=0.6 , Fig. 14 ( f ) there is a favoring of short-range correlations. Similarly, there is a perfect superposition between the critical fields and the post-critical peaks ( \alpha=0.4717 ) and the the pics before the phase transition increase more slowly ( \alpha=0.4071 ), however unlike the previous case they exhibit similar behavior with an increase as the chain size increases, but with a slightly smaller slope for the pre-critical case, which remains higher than the others regardless of N . This demonstrates that short-range correlations impose a chain-length dependency. The scaling factor N^{\alpha} is crucial for understanding how the yield of a quantum thermal engine behaves as the system size increases. The goal is to identify conditions under which increasing N (number of spins or system size) leads to a significant improvement in yield compared to the energy cost. When \alpha>0 , it means that increasing N results in a superlinear growth of cycle work and engine efficiency, which is a sought after goal to achieve better performance on a large scale. Near a critical point, energy fluctuations and heat capacity can diverge, leading to faster yield growth (higher value of \alpha ). This is where critical engines can potentially achieve efficiency close to the Carnot point without sacrificing power. For lower values of \alpha (typically far from the critical point), the yield per spin reaches saturation, as spin interactions are not coherent enough to benefit from increasing system size. In conclusion, the results show that the dynamics of thermal efficiency and work per spin depend strongly on the value of \alpha , which defines the range of interactions between spins. Temperatures ( \beta_{h} and \beta_{c} ) also influence the nature of transitions, with marked differences in spin thermal sensitivity at different values of h_{i} . The observed critical points are generally in agreement with theoretical predictions, although some deviations suggest significant correlation effects in these regimes. We generally observe better performance when the thermal gaps are very large for \alpha=1.2 and \alpha=0.55 . When short-range correlations are favored, we see a more linear critical behavior that depends more on N , particularly with the phase transition. For lower \alpha , there is less dependence on the chain length, which can experience a drop of \pi/N for long-range drops ( \alpha=0.55 ), or similar behavior before and after the phase transition for large N ( \alpha=0.2 ). In conclusion, the performance of quantum thermal machines, whether as engines or refrigerators, is intricately linked to the scaling behavior governed by \alpha . High values of \alpha in engines suggest promising avenues for enhancing efficiency through system size, particularly near critical points where spin correlations are maximized. However, refrigerators, even with intermediate values \alpha , show limited improvement, pointing to the need for optimizing interaction dynamics and minimizing dissipation to achieve better performance. Understanding these scaling laws and their implications can guide the design and optimization of next-generation quantum thermal devices, tailored for specific operational needs. V Conclusion In this study, we investigated the performance of a long-range Ising chain quantum Otto engine and refrigerator under ideal conditions, focusing on the effects of system size N , the interaction range parameter \alpha , and the temperature ratios \beta_{h}/\beta_{c} . Our analysis demonstrates that thermal correlations significantly influence the operational efficiency of these quantum machines, particularly near the critical points of phase transitions. We examined the behavior of all types of correlations for a long chain (N=100). For the quantum heat engine, long-range correlations (small \alpha ) enhance performance in the precritical region, while short-range correlations (larger \alpha ) dominate in the post-critical region. Similarly, for the refrigerator, long-range correlations are also favored in the post-critical field, especially in terms of cooling capacity Q_{c}/N and efficiency \eta_{R} . We then fixed \alpha and varied N to investigate the influence of chain size on performance. For the quantum refrigerator, our results indicate that the cooling efficiency is maximized for smaller system sizes before the phase transition. At the critical point, the behavior of the efficiency \eta_{R} strongly depends on the value of \alpha . For long-range correlations, efficiency decreases for short chains but increases for long chains. In contrast, for short-range correlations, \eta_{R} increases, showing better performance for short chains, while long chains exhibit greater stability and reduced dissipation. For the quantum heat engine, we observed relatively stable performance with minimal efficiency loss when the thermal gap is close and transitions occur before the phase transition, where long-range and intermediate correlations show better performance, even as the system size increases. This stability is attributed to the absence of non-adiabatic effects and the slow, quasi-static nature of the transformation. Additionally, the yield \eta differs from that of the refrigerator, as it depends on the chain size for all values of \alpha . Overall, this study highlights the importance of distinguishing between the critical phase and the diamagnetic phase in understanding the performance of quantum heat engines and refrigerators. The ferromagnetic phase, characterized by significant fluctuations and superior performance, contrasts with the diamagnetic phase, which occurs beyond the critical point and exhibits greater stability but inferior performance. These findings underscore the need for a deep understanding of both static and dynamic conditions to effectively optimize the operation of quantum thermal machines. Appendix A Dynamic of the system A stationary state is reached in the long-time limit t\to\infty . In the following, we will characterize this stationary state through observables. We start with the adjoint version of our master equation, which can be written as: \frac{dO_{H}(t)}{dt}=i\{H,O_{H}(t)\}_{-}+\sum_{n,k}\gamma_{n,k}\Big([1-\zeta f_{n}(\omega_{k})]\\ \times[2b_{k}^{\dagger}O_{H}(t)b_{k}-\{b_{k}^{\dagger}b_{k},O_{H}(t)\}_{+}]\\ +f_{n}(\omega_{k})[2b_{k}O_{H}(t)b_{k}^{\dagger}-\{b_{k}b_{k}^{\dagger},O_{H}(t)\}_{+}]\Big), (9) where O_{H}(t) denotes the Heisenberg form of a Schrödinger observable O . \displaystyle\frac{d}{dt}\langle b_{k}^{\dagger}b_{k}\rangle \displaystyle=-2\sum_{n}\gamma_{n,k}\langle b_{k}^{\dagger}b_{k}\rangle+2\sum_{n}\gamma_{n,k}f_{n}(\omega_{k}), (64a) \displaystyle\frac{d}{dt}\langle b_{k}^{\dagger}b_{q}\rangle \displaystyle=\Big(i(\tilde{\omega}_{k}-\tilde{\omega}_{q})-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})\Big)\langle b_{k}^{\dagger}b_{q}\rangle, (64b) \displaystyle\frac{d}{dt}\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle \displaystyle=\Big(i(\tilde{\omega}_{k}+\tilde{\omega}_{q})-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})\Big)\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle, (64c) \displaystyle\frac{d}{dt}\langle b_{q}b_{k}\rangle \displaystyle=\Big(-i(\tilde{\omega}_{k}+\tilde{\omega}_{q})-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})\Big)\langle b_{q}b_{k}\rangle. (64d) Developing the expressions, we find: \displaystyle\langle b_{k}^{\dagger}b_{k}\rangle(t) \displaystyle=\frac{\sum_{n}\gamma_{n,k}f_{n}(\omega_{k})}{\sum_{n}\gamma_{n,k}}\Big(1-\exp\big(-2\sum_{n}\gamma_{n,k}t\big)\Big) \displaystyle\quad+\langle b_{k}^{\dagger}b_{k}\rangle_{0}\exp\big(-2\sum_{n}\gamma_{n,k}t\big), (65a) \displaystyle\langle b_{k}^{\dagger}b_{q}\rangle(t) \displaystyle=\langle b_{k}^{\dagger}b_{q}\rangle_{0}\exp\Big(i(\tilde{\omega}_{k}-\tilde{\omega}_{q})t-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})t\Big), (65b) \displaystyle\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle(t) \displaystyle=\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle_{0}\exp\Big(i(\tilde{\omega}_{k}+\tilde{\omega}_{q})t-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})t\Big), (65c) \displaystyle\langle b_{q}b_{k}\rangle(t) \displaystyle=\langle b_{q}b_{k}\rangle_{0}\exp\Big(-i(\tilde{\omega}_{k}+\tilde{\omega}_{q})t-\sum_{n}(\gamma_{n,k}+\gamma_{n,q})t\Big), (65d) where the subscript \langle\cdot\rangle_{0} indicates the expectation value at time t=0 . If \sum_{n}\gamma_{n,k}\neq 0 , the limit t\to\infty gives: \displaystyle\langle b_{k}^{\dagger}b_{q}\rangle_{s} \displaystyle=\delta_{kq}\frac{\sum_{n}\gamma_{n,k}f_{n}(\omega_{k})}{\sum_{n}\gamma_{n,k}},\quad\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle_{s}=\langle b_{q}b_{k}\rangle_{s}=0. (66) The equations (65) cease to be valid if, for some pair (k,q) , \gamma_{n,k}=\gamma_{n,q}=0 for all n . In that case, we find: \displaystyle\langle b_{k}^{\dagger}b_{q}\rangle(t) \displaystyle=\langle b_{k}^{\dagger}b_{q}\rangle_{0}e^{i(\tilde{\omega}_{k}-\tilde{\omega}_{q})t}, (67a) \displaystyle\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle(t) \displaystyle=\langle b_{k}^{\dagger}b_{q}^{\dagger}\rangle_{0}e^{i(\tilde{\omega}_{k}+\tilde{\omega}_{q})t}. (67b) Appendix B Adiabatic Transformation The time integral can be evaluated (assuming t is large) as: \int_{0}^{t}e^{\frac{i}{\hbar}\Delta_{mn}t^{\prime}}\,dt^{\prime}\approx\frac{\hbar}{\Delta_{mn}}\left[e^{\frac{i}{\hbar}\Delta_{mn}t}-1\right]\approx\frac{2\hbar}{\Delta_{mn}}\sin\left(\frac{\Delta_{mn}t}{2\hbar}\right) (10) The transition amplitude becomes: a_{n\to m}(t)\approx-\frac{2i}{\Delta_{mn}}\langle m(\lambda)|\frac{dH(\lambda)}{d\lambda}|n(\lambda)\rangle\frac{d\lambda}{dt}\sin\left(\frac{\Delta_{mn}t}{2\hbar}\right) (11) The transition probability P_{n\to m} is the square of the transition amplitude: \displaystyle P_{n\to m}(t) \displaystyle=|a_{n\to m}(t)|^{2}\approx\left(\frac{2}{\Delta_{mn}}\right)^{2}\left|\langle m(\lambda)\middle|\frac{dH(\lambda)}{d\lambda}\middle|n(\lambda)\rangle\right|^{2} \displaystyle\quad\times\left(\frac{d\lambda}{dt}\right)^{2}\sin^{2}\left(\frac{\Delta_{mn}t}{2\hbar}\right) (12) For small transitions, the oscillating term \sin^{2}\left(\frac{\Delta_{mn}t}{2\hbar}\right) , which oscillates between 0 and 1, can be ignored. Thus, the transition probability simplifies to: P_{n\to m}\propto\left(\frac{1}{\Delta_{mn}^{2}}\right)\left|\langle m(\lambda)|\frac{dH(\lambda)}{d\lambda}|n(\lambda)\rangle\right|^{2}\left(\frac{d\lambda}{dt}\right)^{2} (13) We have demonstrated that the transition probability P_{n\to m} is proportional to: P_{n\to m}\propto\left(\frac{1}{\Delta_{mn}^{2}}\right)\left|\langle m(\lambda)|\frac{dH(\lambda)}{d\lambda}|n(\lambda)\rangle\right|^{2}\left(\frac{d\lambda}{dt}\right)^{2} (14) This result forms the basis of the adiabatic condition, which states that for a process to be adiabatic, the rate of change \frac{d\lambda}{dt} must be much smaller than the square of the energy difference \Delta_{mn}^{2} , thereby ensuring that the transition probability remains low. Appendix C Non perfect Thermalization In our paper, we worked on an ideal cycle with perfect ideal and adiabatic transformations, assuming the system is in contact with the two reservoirs for a time t