Thermodynamic Evaluation of Low-GWP A1 Refrigerants for Ultra-Low Temperature Refrigeration Applications

Abstract

Slow market development has caused the lack of low-GWP A1 refrigerants for ultra-low temperature (ULT) refrigeration. Consequently, the high-GWP refrigerant R23 (GWP = 14,600) remains widely used within the ULT sector. For this reason, this paper proposes a comprehensive thermodynamic analysis of recently developed CO₂-based mixtures as low-GWP A1 alternatives to R23 for ULT applications, R469A (GWP = 1357), R472B (GWP = 526), R472A (GWP = 353), and R473A (GWP = 1830). In addition, three system configurations are analysed, the basic two-stage cascade system, and configurations incorporating internal heat exchangers (IHXs). Keeping a constant high-temperature stage (HTS) condensation temperature at 35 °C, three low-temperature stage (LTS) evaporation temperatures are considered, −70, −60, and −50 °C. The highest coefficient of performance (COP) is reached by R23 across all operating conditions and configurations. Among the alternative refrigerants, R473A exhibits the highest COP (0.74% to 1.26% lower than R23). The implementation of IHX results in a reduced COP compared to the basic cycle. R472B is the refrigerant least negatively affected by the IHX implementation due to its high glide. Finally, the environmental impact of R23 is notably reduced by all the alternative refrigerants (up to 95%). This paper’s findings highlight the potential of alternative refrigerants as replacements for R23 in ULT applications.

1. Introduction

The current development of the refrigeration industry focuses on the search for low-global-warming-potential (GWP) refrigerants to mitigate the environmental impact caused by this sector (according to the International Institute of Refrigeration, it is responsible for around 7.8% of global greenhouse gas emissions [1]). Thus, the rising environmental concerns have caused the establishment of international regulations limiting the use of refrigerants depending on their GWP. However, systems used for delivering temperatures below −50 °C are not included in these regulations.

Refrigeration below −50 °C and down to −100 °C is classified as ultra-low temperature (ULT) refrigeration [2], and it is used in pharmaceutical and medical industries to preserve biological samples, vaccines, and other products. It is also used in the chemical and petroleum industries, among others. Vapour compression is the most widely used technology in ULT systems, with the two-stage cascade configuration being the most common architecture [3]. It consists of two independent vapour compression systems (usually working with two different refrigerants), which are thermally connected through the cascade heat exchanger. This heat exchanger serves as both the evaporator for the high-temperature stage (HTS) and the condenser for the low-temperature stage (LTS). Although three-stage cascade systems have also been proposed for ultra-low-temperature refrigeration, their range of applications is usually limited to those requiring temperatures below −120 °C [4,5,6].

Although most of the ULT refrigeration studies found in the open literature analyse theoretically the basic two-stage cascade configuration (incorporating the four essential vapour system’s components: compressor, expansion device, evaporator, and condenser), the experimental works incorporate an additional component to the cascade system, an internal heat exchanger (IHX). It is placed at the LTS, and its function is to increase the LTS compressor suction temperature. Most of the commonly used compressor oils, such as POE 22 or POE 32, which are compatible with HFC, HFO, and HC refrigerants, have a pour point of −40 °C [7]. Thus, the oil properties are degraded below this temperature, causing improper behaviour that can damage the compressor. Several experimental studies that include an IHX at the LTS, connecting the LTS compressor’s suction and discharge, have been found [8,9,10,11]. Furthermore, other experimental studies incorporate an IHX connecting the LTS evaporator outlet with the LTS condenser outlet [12,13]. Despite the use of the IHX in most practical applications, no theoretical studies analysing the operation of a ULT two-stage cascade system with IHX have been found.

Regarding the ULT refrigerants, several alternatives have been proposed to replace the commonly used R404A/R23 and R404A/R508A due to their high GWP (3943, 14,600, and 11,939, respectively). In that way, Sun et al. [14,15] theoretically analysed R404A, R32, R1234yf, R1234ze(E), R161, R1270, R290 and R717 in the HTS and R23, R41, and R170 in the LTS of a two-stage cascade refrigeration system, concluding that R161/R41 exhibit the highest COP compared to the other refrigerant pairs. Similarly, Aktemur et al. [16] theoretically compared R1243zf, R423A, R601, R601A, R1233zd(E), and RE170 for the HTS, being R41 the LTS refrigerant. They concluded that the RE170/R41 has the better performance among the studied refrigerants. Ji et al. [17] performed a theoretical 4E analysis of a ULT cascade system with R290/R170, R290/R1150, R717/R170, R717/R1150, R1270/R170, R1234yf/R170, and R1234yf/R1150. They compared these refrigerant pairs with the HFC baseline R404A/R508B, showing that R290/R170 is the optimal refrigerant pair. It could improve COP by 5.94% and reduce CO2 emissions by 29.67%. Regarding experimental studies, few papers studying alternative refrigerants have been found. Liu et al. [18] and Liu et al. [19] tested R290/R170 in a ULT cascade refrigeration unit. Also, Rodríguez-Criado et al. [20] retrofitted a packaged R290 unit with R170, obtaining a COP between 0.6 and 1.6. Udroiu et al. [13] experimentally compared R290/R170 with R404/R23, showing that R290/R170 offered the highest COP and a 50% emissions reduction.

Although the abovementioned alternative refrigerants offer promising operational results, they have a safety classification of A3 [21], meaning that they are highly flammable. Thus, their use is limited to specific ULT applications (those requiring reduced refrigerant charges, for example). Regarding alternative A1 (non-flammable under standard tests) refrigerants for the HTS, R448A has been proven to be a suitable substitute for R404A in commercial refrigeration [22,23]. Furthermore, Giménez-Prades et al. [11] experimentally analysed R448A as an alternative to R404A in a ULT freezer, with R23 as the LTS refrigerant, concluding that the highest COP was reached by R448A/R23 (up to 10.2% higher than R404A/R23). As for LTS A1 low-GWP alternatives, several CO₂-based mixtures have been developed in recent years. These are R469A, R472B, R472A, and R473A. Despite their recent development, they have already undergone standard safety tests and have been assigned a safety classification of A1 by the ASHRAE [24,25,26,27]. Their thermodynamic characteristics make them suitable for producing temperatures down to −70 °C. Three of them (R469A, R472A, and R473A) have been theoretically analysed by Mota-Babiloni et al. [28]. However, with most of the assumed conditions and operational parameters, results with these mixtures could not be obtained.

As seen, there is a lack of A1 alternative refrigerants to replace the high-GWP R23 and R508A in the LTS of ULT cascade systems. Moreover, theoretical studies analysing different refrigerants do not consider practical aspects, such as the proper lubricant behaviour through an increased LTS compressor suction temperature. For this purpose, this paper proposes a theoretical analysis of R469A, R472B, R472A, and R473A as replacements for R23 in the LTS of a two-stage cascade refrigeration system, using R448A in the HTS. Moreover, two configurations, including internal heat exchangers, are compared to a basic cycle. Three LTS evaporation temperatures are considered (−70 °C, −60 °C, and −50 °C), fixing the HTS condensation temperature at 35 °C. For each evaporation temperature, the cascade heat exchanger temperature (HTS evaporation temperature and LTS condensation temperature) is optimised to deliver the highest energy efficiency. Thus, operational pressures, cascade temperatures, compressor power consumption, compressor discharge temperatures, and COP, among other parameters, are comprehensively studied. Finally, a carbon footprint assessment using the total equivalent warming impact (TEWI) methodology is used to compare the systems’ environmental impact for the different operational conditions, configurations, and refrigerants.

The findings of this study represent an initial approach toward the implementation of the proposed low-GWP alternative refrigerants in real ULT refrigeration systems. Although practical aspects have been considered in the simulations (such as maximum discharge temperature, minimum suction temperature, freezing point, etc.), future experimental research and validation are essential for the design and optimisation of systems operating with new refrigerants.

2. Materials and Methods

This section presents the configurations and refrigerants on which this article is based, as well as the strategy employed, from the assumptions to the final model.

2.1. Configurations

Three two-stage cascade-based cycles are comprehensively analysed in this paper. The first one consists of a basic two-stage cascade system, in which the vapour compression cycles contain the four essential components (compressor, evaporator, condenser, and expansion device). In addition, two other configurations that are commonly used in commercial ULT systems are studied. They include an internal heat exchanger (IHX) whose function is to increase the compressor suction temperature, ensuring proper lubricant behaviour. Thus, the implementation of a compressor suction-discharge IHX (SDHX) and a compressor suction-liquid line IHX (SLHX) is considered. The schematic diagrams of the studied configurations and the corresponding P-h diagrams are shown in Figure 1.

2.2. Refrigerants

This study assesses the performance of several refrigerants for ULT refrigeration applications. R23 is here selected as the low-temperature stage (LTS) baseline refrigerant, and R469A, R472B, R472A, and R473A are analysed as low-GWP LTS alternatives. In all cases, the high-temperature stage (HTS) refrigerant is R448A. Their main properties are presented in

Table 1. Main properties of the studied refrigerants.

Parameter	R448A	R23	R469A	R472B	R472A	R473A
Stage	HTS	LTS	LTS	LTS	LTS	LTS
Composition (wt%)	R32/R125/R134a/R1234yf/R1234ze(E) (26/26/21/20/7)	R23 (100)	R744/R32/R125 (35/32.5/32.5)	R744/R134a/R32 (58/32/10)	R744/R134a/R32 (69/19/12)	R744/R1132a/R23/R125 (60/20/10/10)
Molecular weight (g·mol−1)	86.28	70.01	59.14	54.83	50.39	47.23
Critical temperature (°C)	82.68	26.14	57.05	57.93	50.23	29.88
Critical pressure (MPa)	4.59	4.83	6.60	7.71	7.75	6.96
Normal boiling point (NBP) 1 (°C)	−46.11	−82.25	−78.73	−83.08	−84.52	<−78.40
Triple point (°C)	−150.27	−155.20	−86.00 3	−85.20	−82.45	−78.40
Glide at 1 bar abs (°C)	6.17	0	16.98	28.10	22.77	0.39
Liquid density 1 (kg·m−3)	1350.73	1446.49	1364.01	1352.42	1321.37	1262.99
Vapour density 1 (kg·m−3)	4.64	4.60	3.47	3.09	2.93	6.01
Liquid isobaric heat capacity 1 (kJ·kg−1·K−1)	1.30	1.23	1.54	1.66	1.75	1.81
Vapour isobaric heat capacity 1 (kJ·kg−1·K−1)	0.78	0.69	0.76	0.77	0.77	0.80
ODP	0	0	0	0	0	0
GWP100-years 2	1273	14,600	1357	526	353	1830
ASHRAE st. 34 safety classification	A1	A1	A1	A1	A1	A1

¹ For pressure = 1 bar abs at saturated liquid conditions. For R473A, pressure = 2 bar abs at saturated liquid conditions. ² Extracted from IPCC—AR5. ³ Provided by manufacturer [24].

. Unless indicated, the properties are extracted using REFPROP v10.0 [29].

Regarding the LTS refrigerants’ properties, R23 and R473A exhibit a lower critical temperature compared to R469A, R472B, and R472A. The normal boiling point (NBP) is relatively similar for all the LTS refrigerants. It should be noted that the triple point of R473A (−78.40 °C) is higher than its NBP at 1 bar, so the latter cannot be calculated using REFPROP v10.0. For this reason, the rest of the properties of R473A are obtained at 2 bar instead of 1 bar. Moving to other properties, the glide of R23 (0 K) and R473A (0.39 K) is much lower than that of R469A (16.98 K), R472B (28.10 K), and R472A (22.77 K). Regarding the environmental characteristics, alternative refrigerants lower the GWP by at least 87.46% compared to R23. In addition, all the studied refrigerants have 0 ODP and a safety classification of A1.

2.3. Model and Assumptions

Several working conditions have been selected to simulate the realistic operation of a two-stage cascade ULT refrigeration system. In that way, the average condensation temperature of the HTS is fixed at 35 °C. Three average LTS evaporation temperatures are considered, −70 °C, −60 °C, and −50 °C. Regarding the cascade heat exchanger (CHX) temperatures (condensation temperature of LTS and evaporation temperature of HTS), several values are considered to analyse their influence on different operating parameters. In any case, a temperature difference of 5 K is fixed between the average LTS condensation temperature and the average HTS evaporation temperature [30]. In order to take into account the temperature glide of refrigerants, the average evaporation and condensation temperatures are obtained as suggested by Honeywell [31], Equations (1) and (2), where $T_{bubble}$ corresponds to the saturated liquid temperature and $T_{dew}$ corresponds to the saturated vapour temperature for a given pressure.

$$T_o={\displaystyle \frac{1}{3}}{T}_{bubble}+{\displaystyle \frac{2}{3}}{T}_{dew}$$

(1)

$$T_k={\displaystyle \frac{1}{2}}{T}_{bubble}+{\displaystyle \frac{1}{2}}{T}_{dew}$$

(2)

Moreover, the following assumptions are taken to complete the thermodynamic simulation of the ULT system:

• All system components are assumed to be in a steady-state and steady-flow process [30].
• Target cooling capacity at the LTS evaporator is fixed at 1 kW.
• Superheating degree at the evaporator is 7 K.
• Subcooling degree at the condenser is 2 K.
• Isenthalpic process at the expansion valve [32].
• Compressor isentropic and volumetric efficiencies are a function of the compression ratio [32].
• Eletromechanical efficiency is 0.57 for the LTS compressor and 0.63 for the HTS compressor [13].
• Pressure and heat losses in heat exchangers and connection pipes are neglected [33].

Based on the working conditions and assumptions mentioned, the thermodynamic states at various points in the refrigeration system are calculated using REFPROP v10.0. Thus, the working pressures can be obtained from the corresponding evaporation or condensation temperature and the vapour quality, Equations (3) and (4).

$$P_o=f(x_v=0.66, T_o)$$

(3)

$$P_k=f(x_v=0.5,T_k)$$

(4)

The LTS refrigerant mass flow rate (${\dot{m}}_{ref,LTS}$) is calculated using the proposed cooling capacity (${\dot{Q}}_{o,LTS}$) and the enthalpy lift at the LTS evaporator ($h_{o,LTS,out}-h_{o,LTS,in}$), Equation (5).

$${\dot{m}}_{ref,LTS}={\displaystyle \frac{{\dot{Q}}_{o,LTS}}{(h_{o,LTS,out}-h_{o,LTS,in})}}$$

(5)

The HTS refrigerant mass flow rate (${\dot{m}}_{ref,HTS}$) is obtained from a heat balance at the CHX, Equation (6).

$${\dot{Q}}_{CHX}={\dot{m}}_{ref,HTS}(h_{o,HTS,out}-h_{o,HTS,in})={\dot{m}}_{ref,LTS}(h_{k,LTS,in}-h_{k,LTS,out})$$

(6)

The HTS condenser heat transfer can be calculated with the HTS refrigerant mass flow rate and the enthalpy lift at the HTS condenser ($h_{k,HTS,in}-h_{k,HTS,out}$), Equation (7).

$${\dot{Q}}_{k,HTS}={\dot{m}}_{ref,HTS}(h_{k,HTS,in}-h_{k,HTS,out})$$

(7)

The volumetric flow rate ($\dot{V}$) at compressor suction is calculated using the refrigerant mass flow rate, the density (${\rho }_{suc}$), and the compressor volumetric efficiency (${\eta }_{vol}$), Equation (8). As the analysed alternative refrigerants have been recently developed, there are no available data regarding compression performance, so a generic compressor volumetric efficiency depending on the compression ratio is used [34], Equation (9). The compression ratio is defined as shown in Equation (10).

$$\dot{V}={\displaystyle \frac{{\dot{m}}_{ref}}{{\rho }_{suc} {\eta }_{vol}}}$$

(8)

$${\eta }_{vol}=1-0.06\left({CR}^{{\displaystyle \frac{1}{1.1}}}-1\right)$$

(9)

$$CR={\displaystyle \frac{P_k}{P_o}}$$

(10)

The compressor power consumption (${\dot{W}}_c$) is obtained using the refrigerant mass flow rate, the enthalpy lift ($h_{disch}-h_{suc}$), and the electromechanical efficiency (${\eta }_{em}$), Equation (11).

$${\dot{W}}_c={\displaystyle \frac{{\dot{m}}_{ref}\left(h_{disch}-h_{suc}\right)}{{\eta }_{em}}}$$

(11)

For calculating the compressor discharge enthalpy ($h_{disch}$), the compressor isentropic efficiency, the compressor suction enthalpy, and the compressor isentropic discharge enthalpy are used, Equation (12).

$$h_{disch}={\displaystyle \frac{h_{is,disch}-h_{suc}}{{\eta }_{is}}}+h_{suc}$$

(12)

As happens with the compressor volumetric efficiency, a generic compressor isentropic efficiency, depending on the compression ratio, is used [34], Equation (13).

$${\eta }_{is}=1-0.06\left({CR}^{{\displaystyle \frac{1}{1.1}}}-1\right)$$

(13)

For the configuration incorporating the LTS compressor suction-discharge internal heat exchanger (SDHX), the modified thermodynamic states are calculated by conducting a heat balance on this exchanger, Equation (14).

$$h_{suc}-h_{o,out}=h_{disch}-h_{k,in}$$

(14)

The same is carried out to calculate the modified thermodynamic state when using the suction-liquid internal heat exchanger (SLHX), Equation (15).

$$h_{suc}-h_{o,out}=h_{k,out}-h_{exv,in}$$

(15)

The internal heat exchanger efficiency (${ϵ}_{IHX}$) is calculated as the ratio between the actual heat transfer and the maximum heat transfer at a given set of operating conditions, Equation (16).

$${ϵ}_{IHX}={\displaystyle \frac{h_{c,out}-h_{c,in}}{h_{h,in}-h_{c,in}}}$$

(16)

Finally, the overall efficiency of the refrigeration system, here indicated as the coefficient of performance ($COP$), is calculated as the ratio between the LTS cooling capacity and the total compressor power consumption, Equation (17).

$$COP={\displaystyle \frac{{\dot{Q}}_o}{{\dot{W}}_{c,LTS}+{\dot{W}}_{c,HTS}}}$$

(17)

2.4. Carbon Footprint Assessment

The development of alternative LTS refrigerants responds to the pressing need to reduce the environmental impact of the traditionally used A1 refrigerants such as R23. Thus, this work includes a carbon footprint assessment of the considered ULT refrigeration system working with the proposed refrigerants. The total equivalent warming impact ($TEWI$) metric is used, Equation (18).

$$\begin{array}{cc}\hfill TEWI=& \left(GW{P}_{HTS}{m}_{HTS}+GW{P}_{LTS} m_{LTS}\right) L_{annual} n\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+\left(GW{P}_{HTS}{m}_{HTS}+GW{P}_{LTS} m_{LTS}\right)\left(1-\alpha \right)+n E_{annual} \beta \hfill \end{array}$$

(18)

The direct greenhouse gas emissions associated with accidental leakages are calculated with the global warming potential ($GWP$), the amount of used refrigerant ($m$), the refrigerant leakage rate ($L_{annual}$), and the number of operating years ($n$). The direct emissions associated with the losses produced while recycling refrigerants are obtained with the $GWP$, the refrigerant amount, and the recycling factor at the end of the system’s lifetime ($1-\alpha$). Finally, the indirect emissions are calculated using the number of operating years, the annual energy consumption ($E_{annual}$), and the carbon intensity factor ($\beta$). The latter depends on the emissions needed to generate the system’s required electricity, which in turn depends on the country where the system is located.

For this work, the TEWI is calculated considering several $\beta$ to quantify the influence of the system’s location, keeping a constant leakage rate of 10%, a recycling factor of 90%, and a lifetime period of 15 years. As for the refrigerant charge, the baseline refrigerants (R448A and R23) are assumed to use a mass charge of 1 kg each. This consideration is based on the experimental results obtained by Udroiu et al. [13], who utilised approximately 1 kg of these refrigerants in a two-stage cascade ULT system to produce the cooling capacity assumed in this work. For the LTS alternative refrigerants, the charge is estimated by considering the differences in refrigerant density and the volumetric flow rate at the compressor.

3. Results and Discussion

This section presents and discusses the main results of the ULT system simulations, focusing on thermodynamic parameters, energy efficiency, and potential limitations to implementing the studied refrigerants and configurations in practical applications.

3.1. Model Validation

Prior to simulating the systems proposed in this study, the theoretical model is validated using experimental data provided by Dopazo et al. [35], obtained from a two-stage cascade refrigeration system designed to produce ultra-low temperatures. The operating parameters and design conditions of the experiment are applied to the model developed in this work in order to compare the experimental results with the calculated ones.

As can be seen in Figure 2a,b, the calculated compressor power consumptions agree with the experimentally obtained values, with a relative error of less than 6% in all cases (the average relative error is −2.50% for LTS and 3.48% for HTS). Furthermore, as shown in Figure 2c, the COP maximum relative error is 2.27%, with an average of 1.40%. The comparison between the calculated and the experimental results is thus considered satisfactory. Hence, the numerical model is validated.

3.2. Basic Two-Stage Cascade Configuration

First, the basic two-stage cascade configuration, without incorporating any internal heat exchanger, is thermodynamically analysed when working with the proposed refrigerants.

3.2.1. Influence of the Cascade Heat Exchanger Temperature

The overall system COP as a function of the LTS evaporation temperature and the LTS condensation temperature for all the considered LTS refrigerants is shown in Figure 3.

As can be seen, the highest COP is reached with high evaporation temperatures in all cases. Also, the LTS condensation temperature strongly affects the resulting COP, as it directly influences the compression ratio of both stages, which in turn is related to the compressor power consumption through its isentropic efficiency. Thus, the optimal LTS condensation temperature decreases with lower evaporation temperatures. It is also observed that R23 reaches the highest COP among the considered refrigerants for a given evaporation temperature. Among the alternative refrigerants, R473A exhibits the highest COP, being slightly lower than that of R23.

Another crucial parameter that is strongly influenced by the cascade heat exchanger temperature is the compressor discharge temperature. Most common compressors manufacturers establish a temperature safety limit, above which the compressor can be damaged, reducing its operating life. The temperature limit is commonly set at 120–130 °C [36,37,38,39]. Consequently, a conservative safety limit of 120 °C is considered in this work. The LTS compressor discharge temperature variation with the LTS evaporation temperature and the LTS condensation temperature is shown in Figure 4. According to the safety limit, the highest represented contour line is 120 °C.

The LTS compressor discharge temperature for the alternative refrigerants is considerably higher than that of R23 for a given LTS evaporation and condensation temperatures. In the case of R469A, R472B, and R472A, the main cause of the higher discharge temperatures is a higher compression ratio, which reduces the compressor isentropic efficiency, as well as a higher compressor suction temperature caused by their higher glide. As for R473A, it has the lowest discharge temperatures among the alternative refrigerants for a given LTS evaporation and condensation temperature. In this case, the R473A compression ratio is similar to that of R23, so the isentropic efficiency is higher than that of the other alternative refrigerants. Still, its discharge temperature is notably higher than that of R23.

3.2.2. COP Optimisation

Due to the high influence of the cascade heat exchanger temperature on COP, the temperature for which the COP is optimal without exceeding the compressor discharge limit temperature is presented in

Table 2. Optimal COP results for the basic two-stage cascade configuration.

Refrigerant	To,LTS (°C)	Tk,LTS (°C)	Po,LTS (bar)	Pk,LTS (bar)	CRLTS (-)	To,in,LTS (°C)	Tdisch,LTS (°C)	COP (-)
R23	−70	−20	1.93	13.95	7.20	−70	58.95	0.52
	−60	−15	3.12	16.27	5.21	−60	41.38	0.70
	−50	−10	4.79	18.85	3.94	−50	32.94	0.90
R469A	−70	−24	0.82	7.67	9.41	−79.06	110.6	0.44
	−60	−17	1.42	9.74	6.85	−68.95	78.36	0.64
	−50	−10	2.34	12.21	5.21	−58.81	63.48	0.86
R472B	−70	−26	1.13	9.36	8.25	−79.19	115.5	0.44
	−60	−16	1.91	12.90	6.76	−69.35	100.1	0.64
	−50	−10	3.04	15.44	5.08	−59.67	81.04	0.85
R472A	−70	−24	1.55	11.95	7.71	−75.70	116.9	0.47
	−60	−15	2.56	15.84	6.20	−65.88	98.76	0.66
	−50	−9	4.01	18.90	4.71	−55.12	80.67	0.87
R473A	−70	−20	2.69	19.06	7.10	−70.11	94.42	0.51
	−60	−14	4.32	22.82	5.29	−60.12	70.42	0.69
	−50	−10	6.61	25.61	3.87	−50.14	52.37	0.90

. Other operating parameters, such as pressures, compression ratio, evaporator inlet temperature, and discharge temperature, are also included.

The LTS condensation temperature that maximises the COP is lower at T_o = −70 °C for R469A, R472B, and R472A compared to R23 and R473A. As mentioned earlier, the compression ratios of R469A, R472B, and R472A are considerably higher than those of R23 and R473A, resulting in higher compressor discharge temperatures. As a result, the LTS condensation temperature, which maximises the COP, tends to be lower for these refrigerants when the evaporation temperature is lower. In all cases, the COP is lower than one, which implies that the required electrical power is greater than the produced cooling power.

It should also be noted that, for T_o = −70 °C, the evaporation pressure of R469A is under 1 bar, meaning that at this temperature level the compressor is working at vacuum conditions. Therefore, systems working with this refrigerant with T_o = −70 °C should have strict leakage control to prevent moisture from entering the refrigeration cycle and disrupting the system’s thermodynamic behaviour. Beyond implementing strict leak detection and control measures, the risks related to operating under vacuum conditions can be further mitigated by minimising the number of mechanical connections and weld joints in the piping layout of the installation. This design approach reduces the potential for leaks and improves the overall system integrity.

Another key parameter that determines the safe operation of the system operating with the alternative refrigerants is the LTS evaporator inlet temperature. This is the point where the temperature is the lowest across the entire refrigeration system. Consequently, it must be ensured that this temperature is always above the triple point temperature for each refrigerant (shown in Table 1). As can be seen, the triple point is not reached by any refrigerant under the considered operating conditions, so there is no risk of dry ice formation.

Regarding the COP, it is graphically represented in Figure 5, along with the LTS and HTS compressor power consumption, the cascade heat exchanger and HTS condenser heat transfer, and the LTS refrigerant volumetric flow rate. The results are presented as the relative difference of the alternative refrigerants’ parameters compared to that of R23, as indicated in Equation (19).

$$\%Paramete{r}_{refrigerant}=\left({\displaystyle \frac{Paramete{r}_{refrigerant}-Paramete{r}_{R23}}{Paramete{r}_{R23}}}\right)100$$

(19)

R23 reaches the highest COP across all the evaporation temperatures. Among the alternative refrigerants, R473A has the highest COP, ranging from −1.26% at T_o = −70 °C to −0.74% at T_o = −50 °C compared to R23. Conversely, R472B has the lowest COP values, ranging from −15.37% at T_o = −70 °C to −6.42% at T_o = −50 °C. R469A and R472A have intermediate COP values, with the latter being the refrigerant with the highest COP (−15.26% to −5.11% and −9.53% to −4.02%, respectively). As the target LTS cooling capacity is the same for all refrigerants and all the evaporation temperatures, the difference in COP responds to the variation of compressors’ power consumption (${\dot{W}}_{c,HTS}$ and ${\dot{W}}_{c,LTS}$). Starting with ${\dot{W}}_{c,HTS}$, it is observed that, especially at T_o = −70 °C, it is notably higher when using R469A (24.49%), R472B (33.33%), and R472A (20.07%) in LTS. This happens because, as the CHX temperature is lower with these refrigerants, the compression ratio of the HTS compressor is higher, resulting in a lower isentropic efficiency and thus in higher power consumption. At T_o = −60 °C and T_o = −50 °C, the differences in CHX temperature are reduced, so differences in ${\dot{W}}_{c,HTS}$are also reduced. Regarding ${\dot{W}}_{c,LTS}$, the differences with respect to R23 are lower than those of the HTS compressor, especially at T_o = −70 °C. At this temperature, R472B and R472A have a lower ${\dot{W}}_{c,LTS}$ (−1.93% and −2.10%, respectively). Even if the LTS compression ratio is higher than that of R23, the lower required mass flow rate reduces power consumption. As for R469A and R473A, they have higher ${\dot{W}}_{c,LTS}$ at this temperature level (9.42% and 2.11%, respectively), mainly due to a higher LTS mass flow rate compared to R472B and R472A. At T_o = −60 °C and T_o = −50 °C, ${\dot{W}}_{c,LTS}$ of the alternative refrigerants relative to that of R23 increases for R469A, R472B, and R472A (up to 9.70%, 12.92%, and 11.52%, respectively). The difference in ${\dot{W}}_{c,LTS}$of R473A compared to R23 is relatively constant across the considered evaporation temperatures (1.32% to 5.69%).

The same factors that cause the differences in HTS compressor power consumption between LTS refrigerants are responsible for the differences in the HTS condenser heat transfer (${\dot{Q}}_{k,HTS}$). At T_o = −70 °C, the lower isentropic efficiency of the HTS compressor when using R469A, R472B, and R472A in the LTS causes a higher discharge temperature, increasing ${\dot{Q}}_{k,HTS}$ by 9.86%, 10.18%, and 5.95%, respectively. At T_o = −60 °C and T_o = −50 °C, as happens with the HTS compressor power consumption, the differences on ${\dot{Q}}_{k,HTS}$ between LTS refrigerants are lower (0.30% to 4.52%). The heat transfer at the CHX (${\dot{Q}}_{CHX}$) is relatively similar for all the refrigerants across the different evaporation temperatures (−0.68% to 3.34%). Finally, the highest differences between refrigerants are found in the LTS volumetric flow rate (${\dot{V}}_{LTS}$). The highest ${\dot{V}}_{LTS}$ is reached by R469A across all evaporation temperatures (79.14% to 120.31%). Conversely, the lowest $\dot{V}$ is reached by R473A (−28.06% to −29.73%). R472B and R472A have intermediate values, ranging from 43.03% to 52.73% and from 12.07% to 15.84%, respectively. Overall, it can be concluded that the alternative refrigerants perform better with higher evaporation temperatures.

3.3. IHX Configurations

The configurations incorporating the two types of internal heat exchangers (SDHX and SLHX) are here analysed. As the main purpose of these IHX is to increase the LTS compressor suction temperature, a minimum temperature of −40 °C is considered at this point of the cycle.

3.3.1. Influence of the Cascade Heat Exchanger and the LTS Compressor Suction Temperature

As with the basic cycle configuration, COP is optimised for each considered evaporation temperature, varying the LTS condensation temperature and the LTS compressor suction temperature. For example, Figure 6 illustrates the impact of these parameters on COP for R23.

3.3.2. COP Optimisation

The optimised COP results for the refrigerants considered in the SDHX and SLHX configurations are presented in

Table 3. Optimal COP results for the SDHX configuration.

Refrigerant	To,LTS (°C)	Tk,LTS (°C)	Po,LTS (bar)	Pk,LTS (bar)	CRLTS (-)	Tsuc,LTS (°C)	To,in,LTS (°C)	Tdisch,LTS (°C)	εSDHX (%)	COP (-)
R23	−70	−21	1.93	13.52	6.98	−40	−70	85.47	15.74	0.48
	−60	−15	3.12	16.27	5.21	−40	−60	57.51	13.50	0.67
	−50	−10	4.79	18.85	3.94	−40	−50	36.53	5.16	0.89
R469A	−70	−27	0.82	6.89	8.45	−40	−79.32	118.22	9.61	0.42
	−60	−17	1.42	9.74	6.85	−40	−68.95	89.16	5.95	0.63
	−50	−10	2.34	12.21	5.21	−38.17	−58.81	63.48	0	0.86
R472B	−70	−27	1.13	9.05	7.98	−40	−79.24	118.10	3.30	0.43
	−60	−16	1.91	12.90	6.76	−36.31	−69.35	100.10	0	0.64
	−50	−10	3.04	15.44	5.08	−26.87	−59.67	81.04	0	0.85
R472A	−70	−26	1.55	11.19	7.22	−40	−75.77	117.00	4.26	0.45
	−60	−15	2.56	15.84	6.20	−37.98	−65.88	98.76	0	0.66
	−50	−9	4.01	18.90	4.71	−28.70	−56.12	80.67	0	0.87
R473A	−70	−22	2.69	17.91	6.67	−40	−70.11	118.40	12.50	0.48
	−60	−15	4.32	22.16	5.13	−40	−60.12	84.42	10.19	0.67
	−50	−10	6.61	25.61	3.87	−40	−50.14	55.89	3.52	0.89

and

Table 4. Optimal COP results for the SLHX configuration.

Refrigerant	To,LTS (°C)	Tk,LTS (°C)	Po,LTS (bar)	Pk,LTS (bar)	CRLTS (-)	Tsuc,LTS (°C)	To,in,LTS (°C)	Texv,in,LTS (°C)	Tdisch,LTS (°C)	εSLHX (%)	COP (-)
R23	−70	−20	1.93	13.95	7.20	−40	−70	−33.13	89.09	9.50	0.51
	−60	−14	3.12	16.76	5.37	−40	−60	−22.39	60.11	5.91	0.69
	−50	−10	4.79	18.85	3.94	−40	−50	−13.53	36.53	1.51	0.90
R469A	−70	−27	0.82	6.89	8.45	−40	−80.02	−46.04	118.22	5.09	0.43
	−60	−17	1.42	9.74	6.84	−40	−69.30	−31.04	89.16	2.47	0.64
	−50	−10	2.34	12.21	5.21	−38.17	−58.81	−19.87	63.48	0	0.86
R472B	−70	−27	1.13	9.05	7.98	−40	−79.37	−38.91	118.10	1.53	0.44
	−60	−16	1.91	12.90	6.76	−36.31	−69.35	−25.48	100.10	0	0.64
	−50	−10	3.04	15.44	5.08	−26.87	−59.67	−19.56	81.04	0	0.85
R472A	−70	−26	1.55	11.19	7.22	−40	−75.87	−35.89	117.00	1.93	0.46
	−60	−15	2.56	15.84	6.20	−37.98	−65.88	−21.83	98.76	0	0.66
	−50	−9	4.01	18.90	4.71	−28.70	−56.12	−15.87	80.67	0	0.87
R473A	−70	−22	2.69	17.91	6.67	−40	−70.12	−33.50	118.40	7.01	0.50
	−60	−15	4.32	22.16	5.13	−40	−60.13	−22.45	84.42	4.30	0.68
	−50	−10	6.61	25.61	3.87	−40	−50.14	−13.33	55.89	1.02	0.89

, respectively.

As can be seen, the maximum COP is reached with the lowest considered LTS compressor suction temperature (−40 °C), as increasing it directly increases the LTS compressor power consumption, without any increase in cooling power or decrease in refrigerant mass flow rate. It should be noted that, with an evaporation temperature of −60 °C, the suction temperature of R472B and R472A is above −40 °C without the use of the SDHX (ε_SDHX = 0) due to their high temperature glide. Therefore, the maximum COP in this case corresponds to that of the basic cycle. The same happens at an evaporation temperature of −50 °C with R469A, R472B, and R472A. On the other hand, low-glide refrigerants (R23 and R473A) require higher efficiency SDHXs to reach the minimum suction temperature. Overall, it can be seen that the LTS condensation temperature is lower than that of the basic cycle. With a higher suction and discharge temperatures, a lower condensation temperature, and thus a lower compression ratio, the LTS compressor isentropic efficiency increases, compensating for the increase in LTS compressor power consumption.

Regarding the SLHX configuration, the LTS condensation temperatures that lead to optimal COP results are similar to those of the SDHX configuration. Moreover, as with the SDHX configuration, the maximum COP is reached when the LTS compressor suction temperature is the lowest, except for R469A (T_o = −50 °C), and R472B and R473A (T_o = −50 °C and T_o = −60 °C). In these cases, the suction temperature is above −40 °C without the use of the IHX (ε_SLHX = 0) due to their high temperature glide. With the SLHX, additional subcooling at the evaporator inlet is produced, reducing the required mass flow rate to achieve the desired cooling capacity. However, the increase in suction temperature causes a higher increase in compressor enthalpy lift, decreasing the COP compared to the basic cycle.

The additional subcooling at the evaporator inlet reduces the minimum temperature reached in the cycle. Nevertheless, it can be appreciated that the triple point is not reached in any case, therefore avoiding the risk of refrigerant solidification. The design of the IHX used in this configuration is critical to avoid refrigerant solidification, as an excessively high efficiency would reduce the evaporator inlet temperature below the triple point of CO₂-based refrigerants. On the other hand, a very low efficiency would lead to excessively low suction temperatures, potentially affecting the lubricant behaviour in the compressor. Therefore, the IHX design must strike an appropriate balance to prevent both phenomena.

Another temperature affected by the use of the SLHX is that of the expansion device inlet. The lowest temperature at this point is reached by R469A (−46.04 °C). If an expansion valve with moving parts is considered as the expansion device, the lower temperatures produced by the SLHX should be taken into account to avoid risks of damage. Finally, the LTS compressor discharge temperatures are similar to those of the SDHX configuration.

The subsequent figures depict the COP, the HTS and LTS compressors’ power consumption, the CHX and the HTS condenser heat transfer, and the LTS compressor volumetric flow rate of the analysed refrigerants with the SDHX and SLHX configurations as their relative difference with respect to the basic cycle configuration. In all cases, results correspond to the optimised COP operating parameters. Results are obtained as indicated in Equation (20).

$$\%Paramete{r}_{SDHX,SLHX}=\left({\displaystyle \frac{Paramete{r}_{SDHX,SLHX}-Paramete{r}_{Basic cycle}}{Paramete{r}_{Basic cycle}}}\right)100$$

(20)

The relative difference in the optimised COP between the IHX configurations and the basic cycle is presented in Figure 7. In all cases, the highest COP is obtained with the basic cycle configuration. Conversely, the lowest COP is obtained with the SDHX configuration (up to −7.11% compared to the basic cycle). SLHX reaches intermediate COP values (up to −3.48% compared to the basic cycle). Furthermore, it can be observed that the COP difference of R472B and R472A is lower with the IHX configurations than that of R23, R469A, and R473A, especially at T_o = −70 °C and T_o = −60 °C. As mentioned before, their higher temperature glide causes a higher evaporator outlet temperature, reducing the required IHX heat transfer to reach the minimum considered suction temperature. Consequently, the difference of IHX configurations with respect to the basic cycle is less notable for R472B and R472A. Furthermore, the highest COP difference between IHX configurations and the basic cycle is obtained with R23 and R473A.

As the results are obtained with a fixed LTS cooling capacity, the COP deviations respond to differences in compressors’ power consumption. HTS and LTS power consumption (${\dot{W}}_{c,HTS}$ and ${\dot{W}}_{c,LTS}$) differences between IHX configurations and the basic cycle are presented in Figure 8. The SDHX configuration reaches higher ${\dot{W}}_{c,HTS}$ compared to the SLHX configuration across all the considered evaporation temperatures for a given refrigerant. This is primarily due to the higher required mass flow rate of HTS refrigerant with SDHX. At T_o = −70 °C, both SDHX and SLHX configurations have a higher ${\dot{W}}_{c,HTS}$ than that of the basic cycle (up to 14.94% and 13.90% respectively), regardless of the refrigerant. At T_o = −60 °C, the only case where ${\dot{W}}_{c,HTS}$ is lower than that of the basic cycle (−2.57%) is with the SLHX configuration and R23. In this case, the lower mass flow rate and the higher isentropic efficiency compensate for the increase in the compressor enthalpy lift. Finally, at T_o = −50 °C, the HTS compressor power consumption is similar to the IHX configurations compared to the basic cycle.

As happens with the HTS, the SDHX configuration reaches higher ${\dot{W}}_{c,LTS}$ compared to the SLHX configuration across all the considered evaporation temperatures for a given refrigerant. Moreover, it can be seen that R23 exhibits the highest difference between the IHX configurations and the basic cycle across all the considered evaporation temperatures. The lower LTS condensation temperatures for IHX configurations compared to those of the basic cycle cause a higher isentropic efficiency, especially at T_o = −70 °C, significantly decreasing ${\dot{W}}_{c,LTS}$. In addition, the SLHX configuration causes a decrease in the LTS mass flow rate, further reducing ${\dot{W}}_{c,LTS}$.

Figure 9 shows the differences in CHX and HTS condenser heat transfer (${\dot{Q}}_{CHX}$ and ${\dot{Q}}_{k,HTS}$) of the IHX configurations and the basic cycle. ${\dot{Q}}_{CHX}$ exhibits the same behaviour as ${\dot{W}}_{c,LTS}$, as both parameters are directly influenced by the LTS refrigerant mass flow rate and isentropic efficiency. However, the differences between IHX configurations and the basic cycle are lower for ${\dot{Q}}_{CHX}$ (−3.64% to 2.52%) compared to ${\dot{W}}_{c,LTS}$ (−10.67% to 8.07%). On the other hand, the increased enthalpy lift at the HTS condenser causes a higher ${\dot{Q}}_{k,HTS}$ with SDHX and SLHX compared to the basic cycle.

Finally, the differences in the LTS refrigerant volumetric flow rate (${\dot{V}}_{LTS}$) of the IHX configurations and the basic cycle are presented in Figure 10. The highest differences (up to 10.59%) are reached by R23 across all evaporation temperatures. Moreover, the lower LTS refrigerant mass flow rate causes lower ${\dot{V}}_{LTS}$ for the SLHX configuration compared to that of SDHX. The only cases where the ${\dot{V}}_{LTS}$ is lower than that of the basic cycle is when using R469A, R472B, and R472A at T_o = −70 °C. With these operating conditions, the higher volumetric efficiency caused by a lower LTS condensing temperature and the lower mass flow rate, especially with the SLHX configuration, leads to a reduced volumetric flow rate.

3.4. Heat Transfer Analysis

In addition to the exchanged thermal power, the design of heat exchangers depends on refrigerant-specific parameters such as pressure drop or heat transfer. These parameters, which are related to the thermodynamic and transport properties, provide useful information for estimating the required size of heat exchangers for different refrigerants under the same conditions. Palm [40] developed several figures of merit (FOMs) to visualise the heat transfer and pressure drop differences.

The FOM for pressure drop in a single-phase turbulent flow is calculated as shown in Equation (21). Here, a lower $FO{M}_{∆p}$ value indicates that a lower pressure drop should be expected. This FOM analysis reveals that pressure drop is influenced by factors beyond just fluid viscosity. It also highlights the greater importance of latent heat of vaporization, as it directly influences the mass flow rate needed to achieve a specific cooling capacity, in conjunction with the fluid’s density.

$$FO{M}_{∆p}={\displaystyle \frac{{\mu }^{1/4}}{\rho h_{fg}^{7/4}}}$$

(21)

The results of the pressure drop for the analysed LTS refrigerants are presented in

Table 5. Pressure drop figures of merit.

FOM	To (°C)	R23	R469A	R472B	R472A	R473A
$FO{M}_{∆p,l}$	−70	2.11	1.24	1.11	1.06	N.A.
	−60	2.25	1.29	1.16	1.11	N.A.
	−50	2.43	1.35	1.21	1.18	N.A.
$FO{M}_{∆p,v}$	−70	154.00	232.50	168.48	127.27	N.A.
	−60	106.53	147.29	110.11	84.91	N.A.
	−50	77.34	98.77	75.71	59.53	N.A.

. Note that the thermodynamic and properties are obtained with REFPROP v10.0 for the saturation pressure corresponding to the mean evaporation temperatures shown in the table. Moreover, it should be noted that results cannot be obtained with R473A, as transport properties of one of its components (R1132a) are not available (N.A.). For the liquid phase, R469A, R472B, and R472B exhibit lower values compared to R23 due to their higher latent heat of vaporisation. The relative differences with respect to R23 remain constant across the different considered temperatures, being around 42% for R469A, 48% for R472B, and 50% for R472A. As for the vapour phase, R472A has lower $FO{M}_{∆p}$ than R23 across all the considered temperatures, with a difference of around 20. In contrast, R469A has higher values than R23, and the relative differences between them are more prominent with low temperatures (51% for T_o = −70 °C to 28% for T_o = −50 °C). This is primarily due to the lower density of R469A at lower temperatures compared to other refrigerants. R472B has intermediate values, ranging from +9.4% at T_o = −70 °C to −2.1% at T_o = −50 °C.

Similarly to the pressure drop, the convective heat transfer is evaluated through the FOM extracted from the Dittus–Boelter equation for single-phase heat transfer in turbulent flow, Equation (22). In this case, higher $FO{M}_{SPHT}$ values indicate better heat transfer. The calculated values for liquid and vapour phases are shown in

Table 6. Single-phase heat transfer figures of merit.

FOM	To (°C)	R23	R469A	R472B	R472A	R473A
$FO{M}_{SPHT,l}$	−70	1.12	0.92	0.91	0.97	N.A.
	−60	1.19	0.97	0.95	1.01	N.A.
	−50	1.28	1.02	1.00	1.06	N.A.
$FO{M}_{SPHT,v}$	−70	0.67	0.50	0.50	0.50	N.A.
	−60	0.73	0.54	0.53	0.53	N.A.
	−50	0.81	0.57	0.57	0.57	N.A.

.

$$FO{M}_{SPHT}={\displaystyle \frac{kP{r}^{0.4}}{{\left(h_{fg}\mu \right)}^{0.8}}}$$

(22)

As can be seen, $FO{M}_{SPHT}$ of R23 is higher than that of the alternative refrigerants for liquid and vapour phases, with the difference ranging from 13% to 22% for the liquid phase and from 24% to 29% for the vapour phase. Despite having a higher thermal conductivity compared to R23, the higher latent heat of vaporisation of the alternative refrigerants causes the lower $FO{M}_{SPHT}$ values, suggesting lower heat transfer in single phase.

Finally, given the importance of heat transfer in boiling for vapour compression refrigeration systems, the FOM presented in Equation (23) evaluates this parameter. It is derived from Cooper’s pool boiling correlation and assumes a surface roughness of 1 μm. High $FO{M}_{pb}$ values indicate a high pool boiling heat transfer. Results for the studied refrigerants are presented in

Table 7. Pool boiling figures of merit.

FOM	To (°C)	R23	R469A	R472B	R472A	R473A
$FO{M}_{pb}$	−70	0.067	0.052	0.059	0.067	0.080
	−60	0.078	0.060	0.068	0.077	0.092
	−50	0.090	0.069	0.077	0.087	0.106

.

$$FO{M}_{pb}=p_r^{0.12}(-{{\mathit{log}}_{10}{p}_r)}^{-0.55}{M}^{-0.5}$$

(23)

R473A exhibits higher $FO{M}_{pb}$ compared to R23, with a relative difference of approximately 18%. In contrast, R469A, R472B, and R472A have lower values than R23, with differences of 23%, 14%, and 2%, respectively. These lower values are primarily caused by their lower reduced pressure.

In conclusion, the deviations in pressure drop and heat transfer obtained through FOMs indicate that the alternative refrigerants have a lower single-phase heat transfer compared to R23. Regarding the pool boiling heat transfer, R473A reaches higher values compared to R23, suggesting an improved heat transfer at the evaporator. The pressure drop of the alternatives is lower than that of R23 for the liquid phase. In the vapour phase, R472A has a lower pressure drop, and R469A has a higher pressure drop than R23, regardless of the temperature, while R472B reaches intermediate values, being higher or lower than R23 depending on the temperature.

3.5. TEWI

An environmental evaluation is conducted following the Total Equivalent Warming Impact (TEWI) methodology. It calculates the total equivalent CO₂ emissions that a system produces throughout its useful life.

Figure 11 shows the equivalent CO₂ emissions reduction in the alternative refrigerants compared to R23 for the optimised COP. The SDHX configuration is used for calculations, and three emission factor levels ($\beta$) are taken to illustrate the impact of electricity generation emissions. These emission factors correspond to Germany (0.329 kgCO₂ kWh⁻¹), EU’s average (0.210 kgCO₂ kWh⁻¹), and France (0.050 kgCO₂ kWh⁻¹). These values are extracted from the EEA database [41].

The reduction in equivalent CO₂ emissions is strongly influenced by the emission factor. As it decreases, the emissions related to refrigerant leakages become more prominent than those related to energy efficiency (COP). Therefore, the alternative refrigerants reduce emissions more significantly due to their lower GWP compared to that of R23. For higher emission factors, energy efficiency strongly influences the resulting equivalent CO₂ emissions. Besides reaching lower emissions reduction with alternative refrigerants, there is a noticeable difference between evaporation temperatures. Taking as an example the average emission factor of the EU, the highest emissions reduction is achieved by R472A at T_o = −50 °C (39.01%) and T_o = −60 °C (33.52%), and by R473A at T_o = −70 °C (24.79%). Conversely, the lowest emissions reductions are achieved by R469A across all the evaporation temperatures (14.23% to 35.39%. In the case of France, these reductions increase up to 69.11%, 67.72%, 63.83%, and 63.76% by R472A, R472B, R473A, and R469A, respectively. Finally, with the emission factor of Germany, the reductions reach maximum values of 29.05%, 26.92%, 28.09%, and 26.01% for the same refrigerants. The emission reductions, depending on the emission factor and the evaporation temperature, are summarised in

Table 8. TEWI percentage reduction in alternative refrigerants compared to R23.

$\mathit{\beta }$ (kgCO2 kWh−1)	To (°C)	R469A	R472B	R472A	R473A
0.210	−70	14.23	18.75	23.27	24.79
	−60	27.48	30.75	33.52	31.19
	−50	35.39	37.04	39.01	36.98
0.050	−70	48.64	53.59	56.43	53.45
	−60	58.25	62.69	64.57	59.38
	−50	63.76	67.62	69.11	63.83
0.329	−70	5.51	9.92	14.87	17.52
	−60	18.49	21.42	24.45	22.96
	−50	26.01	26.92	29.05	28.09

.

4. Conclusions

In this paper, the recently developed ULT mixtures R469A, R472B, R472A, and R473A are evaluated as R23 replacements in the low-temperature stage (LTS) of a two-stage cascade refrigeration system, using R448A as the high-temperature-stage (HTS) refrigerant. The alternative refrigerants improve the environmental impact of R23, lowering the GWP by up to 97.5%, and maintaining the safety classification A1 (low flammability and toxicity). Apart from the basic two-stage cascade configuration, two other configurations, including an internal heat exchanger (IHX), are analysed, the SDHX configuration, in which the IHX connects the LTS compressor suction and discharge, and the SLHX configuration, in which the IHX connects the LTS compressor suction with the liquid line. The implementation of an IHX to the ULT system responds to the practical necessity of increasing the LTS compressor suction temperature to ensure proper oil behaviour.

For each LTS refrigerant and configuration, three LTS evaporation temperatures are considered (−70 °C, −60 °C, and −50 °C), fixing the HTS condensation temperature at 35 °C. For each evaporation temperature, the cascade heat exchanger temperature (HTS evaporation temperature and LTS condensation temperature) is optimised to deliver the highest energy efficiency.

The highest COP is reached by R23 and R473A across all the tested conditions and configurations. R469A, R472B, and R472A exhibit lower values, with a maximum difference of −15.37% compared to R23. Moreover, R473A’s volumetric flow rate at compressor suction is up to 29.73% lower compared to R23, indicating that a smaller compressor is needed to produce the same cooling power. Conversely, the refrigerant with the highest suction volumetric flow rate is R469A, with a difference of up to 120.31% compared to R23.

Regarding the configurations, the better IHX configuration in terms of energy efficiency and volumetric flow rate at the compressor suction is SLHX in all cases. However, given that this IHX produces a higher subcooling at the expansion device inlet, the lower temperatures at this point should be considered in practical applications to ensure proper operation of expansion valves with moving parts. The SDHX is thus the better configuration for ensuring a safe operation. Furthermore, performance indicators are less degraded with the IHX configurations compared to the basic cycle with high glide refrigerants, as less heat transfer at the IHX is needed to reach an adequate LTS compressor suction temperature.

Although this study is conducted theoretically and experimental work is needed to prove the alternative refrigerants’ performance, their potential as low-GWP and A1 replacements for R23 in ultra-low-temperature refrigeration applications is highlighted.