Tribological properties of low-temperature time-dependent pretreated graphite for mechanical seal pairs in high-speed turbopump

The friction coefficient and wear rate of pretreated graphite with liquid nitrogen were obtained by using a ball-on-disk tester, and the wear of GCr15–graphite seal pairs with the low-temperature time-dependent pretreatment was discussed by comparing the wear morphology. The results show that liquid nitrogen pretreatment can affect the hardness and interlayer spacing of graphite. The range of the friction coefficients of pretreated graphite changes from 0.17 to 0.22. With the increase of liquid nitrogen pretreatment time, the wear mechanism of graphite would change from dominated three-body wear to adhesion wear. The experimental results of the mechanical seal with liquid nitrogen pretreatment show that the wear rate of stator is less than 0.00165 mm3·N−1·m−1, and the graphite shows a good low-temperature compatibility.


Introduction
The high-speed turbopump in the liquid rocket engine for deep-space exploration works under extreme conditions such as low temperature, high speed, high pressure, and transient start-up.The shaft-end mechanical seal of the turbopump is a key part for ensuring the safe and reliable operation of the whole pump system [1,2].For meeting the rocket engine sequence working requirements, the seal is directly exposed to the environment of low-temperature fluid medium (liquid hydrogen or liquid oxygen) and has to work continuously in a special period (about 300 s) [3].The special period corresponds to the complete operation time of the cryogenic liquid propulsion rocket, that is, the full sequence working time is from ignition fast start-up to the completion of the given launch mission, and the whole process lasts about 300 s.In this special period, the seal of the turbopump needs to work continuously and stably, including in the rocket ignition fast start-up stage and high-speed stable operation stage, until the low-temperature fluid medium is completely exhausted.The mechanical seal pairs are generally made of graphite and steel, where the rotor of the seal is made of steel, and the stator is made of graphite [4,5].The change of the graphite properties under a low-temperature condition is extremely complex, so the study of graphite properties with different low-temperature pretreatment time is of great practical value for accurately predicting the properties of seal pairs and improving the operational reliability of mechanical seal.
Meanwhile, in order to better understand the tribological properties of graphite under harsh conditions, researchers have conducted a lot of friction and wear tests under different conditions (such as dry friction, high temperature, and water lubricated).For example, Zhao et al. [24] pointed the worn taper angle appeared on the surface of impregnated graphite, and the angle was in the opposite direction at the inner and outer diameters.Zhang et al. [25] investigated graphite wear under the dry friction condition and predicted the relationship between frictional dissipation and wear based on graphite/ WC-Ni pairs.Luo et al. [26] , Kim and Kim [27], and Yan et al. [28] investigated the effect of high temperature on the tribological properties of graphite, and Iwasa et al. [29] studied the effect of low temperature on the tribological properties of graphite and concluded that the friction coefficient was inversely proportional to the material hardness.Goilkar and Hirani [30] found that the friction torque and wear rate can be reduced substantially under a low equilibrium ratio condition.Ding and Jiang [31] investigated the tribological properties of isostatic graphite and carbon-graphite under dry friction and water-lubricated conditions, respectively, and found that the friction coefficient under water lubrication was higher than that under dry friction, while the wear rate was lower, and the tribological properties of both materials were improved under the water lubrication condition.Although so many researchers have conducted the research on the tribological properties of graphite under complex conditions, the study on graphite in a low-temperature condition is still limited.
As previously stated, the mechanical seal in the high-speed turbopump is a core component of the low-temperature liquid rocket engine, and its safety directly affects the work reliability of the whole engine.Some statistics show that the main factor of the seal failure is from the wear loss and undesirable friction state.Because the mechanical seal in the turbopump is in extreme environments for long presoaking time, its wear mechanism in this extreme condition is more complex.In particular, the experimental test study on the low-temperature time-dependent pretreated graphite of the mechanical seal is still blank.In this paper, according to the sequence working characteristics during the start-up stage of the shaft-end mechanical seal, the method of low-temperature liquid nitrogen time-dependent pretreatment of graphite was proposed.The friction and wear properties of GCr15-graphite seal pairs with six different low-temperature time-dependent pretreatment time were obtained, and the wear mechanism of graphite was discussed.In order to verify the tribological properties of graphite in a lowtemperature environment, a special seal experiment test with the liquid nitrogen as the sealed fluid was carried out.The results will provide important experimental references for improving the properties of the mechanical seal pairs in high-speed turbopump.
2 Experimental test

Experimental equipment
The tribological properties of low-temperature time-dependent pretreated graphite and GCr15 alloy structural steel at varied pressure multiplied by velocity Pv values were evaluated by the ball-on-disk module of a multifunctional material surface property tribometer tester (CFT-I), as shown in Fig. 1 [32].Figure 1(a) describes the schematic of the friction coefficient measuring units by using the ball-on-disk module, and the varied loads F (N) are applied to the disk specimen through the ball specimen fixture.The range of Hertzian contact pressure between ball (GCr15 steel) and disk (graphite) is from 0 to 1,400 MPa, and the range of load between ball and disk is from 0 to 50 N.A rotating speed of the disk specimen can be adjusted from 0 to 300 r•min The friction coefficient is obtained directly by the above friction coefficient measuring units, and the wear rate W (mm 3 •N −1 •m −1 ) is obtained by Eq. ( 1): where V is the wear volume loss (mm 3 ), and D is the sliding distance (m), which is the product of time (s) and velocity (m•s −1 ).V is calculated by Eq. (2): where R is the average equivalent rotation radius of the wear scar (mm), which is defined as the rotating radius of the ball, as shown in Fig. 2, and S is the wear scar cross-sectional area (mm 2 ), which is calculated by Eq. ( 3): where L is the wear scar width (mm), a is the wear scar starting position, b is the wear scar end position, and f(x) is the curve profile function of wear scar depth (mm), which is obtained directly by a displacement sensor (No. 10 in Fig. 1).

Experimental material
The disk specimens were made of the special highdensity fine-grained graphite.The diameter of disk specimen is Ø30 mm, while the thickness is 10 mm.The ball specimen material is GCr15, and the diameter is Ø6 mm.The surface roughness R a of the graphite disk is 0.8, which is polished with a 320-mesh grinding disk under the water lubrication condition and dried in a vacuum drying apparatus to remove the surface moisture.The physical properties of the graphite before pretreatment and the GCr15 steel are shown in Table 1.
The graphite specimen was pretreated by soaking in liquid nitrogen (−169 °C) with different time from 1 to 24 h.Based on the time graphite is soaked in liquid nitrogen, the low-temperature time-dependent pretreated graphite is defined.

Experimental design
In order to be consistent with the pretreatment time of the actual sequence working condition of the graphite for the mechanical seal in a low-temperature environment, the graphite is set to soak in liquid nitrogen (−169 °C) for different periods of time.By considering the actual mechanical seal in the liquid rocket engine start-up sequence and possible repeat start-up requirements, the graphite is set specially to  for every working condition.There are total 54 sets of experiments.

Friction coefficient and wear rate
In total, 54 sets of experiments were finished.The results of three friction experiments for every working condition are shown as follows.When the Pv is 7 MPa•m•s −1 , and the pretreatment time is 1 h, the friction coefficients of graphite against GCr15 are shown in Fig. 4(a), the horizontal coordinate is the test time, and the vertical coordinate is the friction coefficient.Each test specimen has been subjected to three repeated tests, and each test lasts for 3 min.As shown in Fig. 4(b), it is clear that the friction coefficient decreases with the increase of Pv values.The reason may be as follows: Under the dry friction condition, the increase of Pv value accelerates the formation of graphite transfer layer on the surface of GCr15, which is beneficial for a lower friction coefficient [14].As an illustration, the special tests of friction coefficients with pretreatment time for  As shown in Fig. 5, under the working condition with a low Pv value, the friction coefficient appears to suddenly decrease and increase, and this phenomenon shows that the graphite transfer film of the contact surface ruptures.With the increase of the Pv value, this phenomenon becomes less obvious.Meanwhile, we tried to observe the wear scar on the ball surface after the test, but the pretreatment can seriously damage the transfer film that may be attached on the ball surface, so the existence of the transfer film cannot be directly observed.
According to Eq. ( 1), the wear rates under different working conditions are calculated and displayed in Fig. 6.
Under 2.4 and 4.7 MPa•m•s −1 working conditions, the graphite shows less fluctuation in wear rate and has a lower wear rate.Under the 7 MPa•m•s −1 working condition, the wear rate fluctuates more, and the wear rate is also higher; the maximum wear rate is when the pretreatment time is 12 h, and the minimum wear rate is when the pretreatment time is 3 h.Explaining the possible reasons, the wear mechanism changes may cause changes in the wear rate, and further discussion of the wear mechanism of graphite under different working conditions will follow based on the wear morphology.

Wear morphology results and analysis
As shown in Figs. 4 and  A metallographic microscope (GX41, Olympus, Japan) has been used to observe the wear morphology.
The observed area, which needs to contain the complete wear scar of each graphite specimen, is properly selected, and the clear wear scar results can be displayed by the metallographic microscope.Figure 7 shows the metallographic micrographs of graphite before and after experiment with pretreatment time for 12 h under 2.4 and 4.7 MPa•m•s −1 working conditions.
Figures 7(a) and 7(b) show the wear morphologies with pretreatment time for 12 h before and after experiment under the 2.4 MPa•m•s −1 working condition, respectively.The rough surface of graphite can be clearly seen in Fig. 7(a), the surface micro-convex body is clearly seen in Fig. 7(b) due to the shear force on the contact surface, and it causes scrapes and furrows on the graphite surface.Figures 7(c) and 7(d) show the wear morphologies before and after experiment under the 4.7 MPa•m•s −1 working condition, respectively.The scrapes and furrows as well as flaking points can be observed in Fig. 7(d), and the wear mechanism is a mixture of three-body wear and adhesive wear.
Figure 8 shows the metallographic micrographs of graphite before and after experiment with pretreatment time for 6 and 12 h under the 7 MPa•m•s −1 working condition.Figures 8(a www.Springer.com/journal/40544| Friction working condition, respectively.In Fig. 8(b), the scrapes and furrows on the wear surface can be clearly seen, the wear mechanism is three-body wear, and the three-body wear is more obvious than that under the lower Pv value condition.Figures 8(c) and 8(d) show the wear morphologies before and after experiment with pretreatment time for 12 h under the 7 MPa•m•s −1 working condition, respectively.As shown in Fig. 8(d), it can be observed that there are obvious flaking points, there is no three-body wear on the surface, and the wear mechanism is adhesive wear.Comparing the two wear behaviors, the reasons for the change of the wear mechanism will be discussed in Section 3.3.

Wear mechanism
In order to understand the relationship between the tribological properties and the surface hardness of graphite (non-metallic material), the surface hardness before and after low-temperature pretreatment with different time was tested by using the Shore hardness tester (HS-19GD, API, China).Six hardness measurements at three different positions on the positive and negative surfaces of the each low-temperature time-dependent pretreated graphite disk specimen are carried out, and the surface hardness is the mean value of six measurements.The variation of hardness with pretreatment time is shown in Fig. 9.

Surface hardness (HSD)
Fig. 9 Surface hardness of graphite variation with pretreatment time.
As shown in Fig. 9, it is clear that the low-temperature liquid nitrogen pretreatment has a significant effect on the graphite surface hardness.The influence of the pretreatment time on the surface hardness of graphite is complex.When the pretreatment time is from 0 to 1 h, the surface hardness of graphite increases; and then when the time changes from 1 to 3 h, the hardness decreases.When the pretreatment time is maintained from 3 to 12 h, the hardness slightly increases as the time.As the pretreatment time is over 12 h, the hardness decreases substantially.The highest surface hardness of graphite is 56.40 HSD And when the surface hardness is the same, the wear changes from three-body wear to adhesive wear as the Pv value increases.Figure 10 shows the wear morphologies of graphite with pretreatment time for 1 and 3 h under working conditions with different Pv values.As shown in Fig. 10, the wear is mainly three-body wear under the working condition with a low Pv value, and the wear is mainly adhesive wear under the working condition with a high Pv value.
With the increase of Pv value, the wear changes from three-body wear to adhesive wear.
According to the study on the effect of helium on graphite properties by Luo et al. [33], the graphite interlayer spacing d 002 increases in helium, and the reduction of van der Waals forces between graphite interlayers will lead to decrease the friction coefficient.The reason may be that as the Pv value increases, the changes in graphite interlayer spacing and surface hardness together affect the tribological properties of graphite, thus causing a change in the wear mechanism.In order to investigate the fact that such phenomena may also be caused by low-temperature liquid nitrogen pretreatment, a set of tests were also carried out in this paper, and the X-ray diffraction (XRD; MiniFlex600, Rigaku, Germany) was used to measure the graphite's interlayer spacing, which may have an effect on its tribological properties.The XRD tests were performed on graphite disks with liquid nitrogen pretreatment time for 0, 1, 3, 6, and 12 h.Figure 11(  As shown in Fig. 11(b), it can be seen that the θ increases first, and then decreases, and finally increases with the increase of pretreatment time.Therefore, through the equation of d 002 = λ/(2sinθ), it can be obtained that the graphite interlayer spacing first decreases, and then increases, and finally decreases with the increase of pretreatment time.The graphite interlayer spacing becomes smaller than that without pretreatment, so the van der Waals forces between graphite interlayers are increased.Combined with Fig. 9, it is found that the graphite surface hardness has a certain relationship with the graphite interlayer spacing.With the decrease of graphite interlayer spacing, the graphite surface hardness According to the results of Figs. and 8(d), it is found that the wear of graphite with liquid nitrogen pretreatment time for 6 h is three-body wear, while the wear for 12 h is adhesive wear.Combined with Figs. 9 and 11, the reason for the change in the wear is explained that the graphite interlayer spacing with pretreatment time for 6 h is higher than that with pretreatment time for 12 h, the surface hardness is lower, and the shedding graphite particles can slip on the contact surface, which leads to the three-body wear and lower wear rate attributed to good self-lubricating properties of graphite.According to Figs. 4(b) and 6, it is found that the graphite with pretreatment time for 3 h has the maximum friction coefficient and the minimum wear rate under the 7 MPa•m•s −1 working condition.Combined with Figs. 9, 10, and 11, the reason can be explained that the wear of graphite with pretreatment time for 3 h is mainly adhesive wear, so the friction coefficient is higher.The surface hardness of graphite with pretreatment time for 3 h is lower, and the graphite interlayer spacing is higher.When the Pv value is high, the shedding graphite layer is deformed and adhered to the contact surface because of the high contact pressure, so the wear rate is lower.
According to the test results, the possible variation of the wear mechanism between the friction pairs of graphite against GCr15 with Pv value and pretreatment time, respectively, is shown in Fig. 12. Figure 12(a) is the three-body wear, Fig. 12(b) is the mixed wear of three-body wear and adhesive wear, Fig. 12(c) is the adhesive wear, and Fig. 12(d) is the three-body wear under a high Pv value.The color of the graphite indicates different pretreatment time.The Pv value increases as the color of the force varies from light brown to brown.The browner the force is, the higher the Pv value is.
Compared with Figs.12(a), 12(b) and 12(c), the wear changes from three-body wear to adhesive wear with the increasing Pv value for the same pretreatment time.Graphite particles, resulted from the rotational movement of the contact surfaces, cause the three-body wear.The high Pv value leads to graphite layer shedding, and further severe wear appears in Figs.12(b) and 12(c).Compared with Figs.12(c) and 12(d), the wear changes from three-body wear to adhesive wear with the increasing pretreatment time for a high Pv value.As shown in Fig. 12(d), the graphite particles are slightly deformed under high pressures because of the reduction of graphite hardness.Combined with Figs. 9 and 11(b), the graphite interlayer spacing with pretreatment time for 6 h is higher than that with pretreatment time for 12 h, and the hardness is lower, so the shedding graphite layer  will act as a lubricant on the friction surface, which can explain that the wear rate of graphite with pretreatment time for 6 h is lower than that with pretreatment time for 12 h at 7 MPa•m•s −1 .

Test of mechanical seal
In Section 3.1, we mainly tested the friction and wear properties of the metal-graphite pairs with liquid nitrogen time-dependent pretreatment under different Pv values.However, the Pv value in the low-temperature simulation experiment of the actual mechanical seal in the high-speed turbopump is a dynamic parameter, and even the Pv value generated by the whole turbopump operation conditions on the seal is also a dynamic change parameter.Hence, in order to better understand the wear properties of graphite materials during the dynamic change process of different Pv values under the actual low-temperature operation conditions, the experimental tests on the mechanical seal (the metal rotor and the graphite stator with the above optimal 3 h liquid nitrogen pretreatment) need be carried out, which are of great engineering value to accurately obtain the performance of the mechanical seal under the harsh low-temperature high-speed operation conditions.
The pretreated graphite was made into a stator of mechanical seal for high-speed turbopump, the experimental test with liquid nitrogen as the sealed fluid was completed.The mechanical seal test system and test have been introduced in Ref. [3].The seal rotating speed is from 0 to 21,000 r•min −1 , which includes the start-up and stable running stages.The closed force of the mechanical seal is 528 N, and the total running time of the mechanical seal is 350 s.The existing references [1,3] and experimental test results show that the wear failure of the seal pairs is the most failure of the mechanical seal, so the lowest wear rate of the seal pairs is the optimal parameter for selecting the seal face material and structure.The above results in Section 3.1 show that the graphite with pretreatment time for 3 h has the lower wear rate, so the graphite with pretreatment for 3 h is selected as the stator material.The graphite in the stator was pretreated in low-temperature liquid nitrogen for 3 h.According to the study in Ref. [34], the stator and rotor of the mechanical seal will be separated at a certain speed.When the closing force is 528 N, the separation occurs when the rotational speed is about 10,000 r•min −1 .The time from 0 to separation speed is about 3 s.The contact force decreases gradually as the speed increases, and when The running distance before the seal separates is the half of the separation speed multiplied by the separation time.
Figure 13 shows the rotor and stator of the mechanical seal before the test, and Fig. 14 shows the wear morphologies after the test.
As shown in Fig. 13(a), the material of the rotor is 9Cr18 stainless steel, whose mechanical and tribological properties are similar to those of GCr15 [35].As shown in Fig. 13(b), the stator is a graphite ring, which is fixed on the foundation bed.The parameters of the rotor and stator are shown in Table 3.
As shown in Fig. 14, there are obvious severe scrapes on the rotor, and the grooves on the surface of the rotor are worn out.The reason for this phenomenon may be the increasing hardness of the stator with liquid nitrogen pretreatment.The stator has obvious flaking points, scrapes, and furrows, and the wear mechanism is a combination of three-body wear and adhesive wear.The reason for the flaking   points on the stator surface may be caused by the two-phase flow of liquid nitrogen [1].According to Ref. [34], the mechanical seal has a separation speed under different closed forces; and when the of mechanical seal is higher than the separation speed, the seal end face separates.Despite the separation of the seal end face, the contact of the seal end face may be caused by two-phase flow corrosion, axial vibration from the other part (such as impeller and floating ring seal), and other factors.The wear rates of the rotor and stator are 0.033 and 0.00165 mm 3 •N −1 •m −1 , respectively.A change in the diameter of the rotor occurs after the test, and this parameter increases by 0.06 mm, which means that severe deformation exists under the low-temperature conditions.The test results further show that the friction status of the rotor and stator is extremely serious for the low-temperature mechanical seal, and the wear is affected by the dry friction in the transient start-up stage, the shock caused by the axial vibration of the pump system and sealed low-temperature fluid, etc. Speaking from the more specific aspect, based on the lower wear rate from the results of Fig. 6, the graphite with 3 h pretreatment is a better choice for this kind of the mechanical seal test.On the other hand, in order to ensure a lower friction coefficient of the seal, as shown in Fig. 4, longer pretreatment time (6 or 12 h) may be more beneficial.The further study on the tribological properties of the graphite with longer pretreatment time still needs to be investigated in the future.

Conclusions
In order to meet the low-temperature, high-speed, −1 , and www.Springer.com/journal/40544| Friction work time of 3 min is set as testing parameter.The schematic of wear rate measuring module is shown in Fig. 1(b).

Fig. 2
Fig. 2 Sketch of wear scar cross section.

Figure 4 (
b) shows the variation of friction coefficients with pretreatment time under different Pv values (2.4,4.7, and 7 MPa•m•s −1 working conditions), the horizontal coordinate is the Pv value, and the vertical coordinate is the friction coefficient.
1 and 6 h under different Pv values are shown in Fig. 5. From Fig. 4(b), the friction coefficient of graphite varies with the same trend under 2.4 and 4.7 MPa•m•s −1

Fig. 6 Fig. 4
Fig.6 Wear rates of graphite with pretreatment under different working conditions.
6, under 2.4 and 4.7 MPa•m•s −1 working conditions with pretreatment time for 12 h, the friction coefficient and wear rate of graphite are optimized, which corresponds to the lowest friction coefficient and lower wear rate.Under the 7 MPa•m•s −1 working condition, the wear rate of graphite with pretreatment time for 12 h shows a sudden increase.Hence, the wear morphologies of graphite with pretreatment time for 12 h under 2.4 and 4.7 MPa•m•s −1 working conditions and with pretreatment time for 6 and 12 h under the 7 MPa•m•s −1 working condition are carried out.

Fig. 8
Fig. 8 Wear morphologies of graphite: (a) before experiment, (b) with pretreatment time for 6 h under 7 MPa•m•s −1 working condition, (c) before experiment, and (d) with pretreatment time for 12 h under 7 MPa•m•s −1 working condition.
a) shows the XRD results, and Fig. 11(b) shows the asymmetric (002) peaks of graphite with five different pretreatment time.The graphite interlayer spacing can be calculated by the equation of d 002 = λ/(2sinθ), where λ is the wavelength of the target material (for Cu, λ = 0.1541874 nm), and θ is the diffraction angle, which can be obtained from Fig. 11(b).

Fig. 12
Fig.12 States of (a) three-body wear, (b) mixed wear of three-body wear and adhesive wear, (c) adhesive wear, and (d) three-body wear under a high Pv value.

Table 1
Physical properties of graphite and GCr15.

Table 2
Experimental design.
www.Springer.com/journal/40544| Friction working conditions.With the increase of pretreatment time, the friction coefficient of graphite is decreasing, and then slightly increasing, and then decreasing and finally increasing slightly.The minimum friction coefficients are 0.213 and 0.197 with pretreatment time for 12 h.The minimum values are 14.8% and 10.5% lower than that of graphite without pretreatment, respectively.For the 7 MPa•m•s −1 working condition, with the increase of pretreatment time, the friction coefficient of graphite is first increasing, and then decreasing, and finally slightly increasing.When the pretreatment time is more than 6 h, the friction coefficient is lower and remains nearly stable at 0.177.The complex variation of friction coefficient may be related to the material surface hardness, which will be discussed in Section 3.3.Fig. 5 Friction coefficients of graphite with pretreatment time for 1 and 6 h under 2.4, 4.7, and 7 MPa•m•s −1 .

Table 3
Parameters of rotor and stator.