Surface heat transfer characteristics of R404A pulsed spray cooling with an expansion-chambered nozzle for laser dermatology

Accepted Manuscript Title: Surface heat transfer characteristics of R404A pulsed spray cooling with an expansion-chambered nozzle for laser dermatology Author: Rui Wang, Zhifu Zhou, Bin Chen, Feilong Bai, Guoxiang Wang PII: DOI: Reference:

S0140-7007(15)00259-5 http://dx.doi.org/doi: 10.1016/j.ijrefrig.2015.08.016 JIJR 3140

To appear in:

International Journal of Refrigeration

Received date: Revised date: Accepted date:

3-4-2015 18-8-2015 19-8-2015

Please cite this article as: Rui Wang, Zhifu Zhou, Bin Chen, Feilong Bai, Guoxiang Wang, Surface heat transfer characteristics of R404A pulsed spray cooling with an expansion-chambered nozzle for laser dermatology, International Journal of Refrigeration (2015), http://dx.doi.org/doi: 10.1016/j.ijrefrig.2015.08.016. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Surface heat transfer characteristics of R404A pulsed spray cooling with an expansion-chambered nozzle for laser dermatology Rui Wanga, Zhifu Zhoua, Bin Chena, *, Feilong Baia, Guoxiang Wanga, b a

State Key Laboratory of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi, China b

Department of Mechanical Engineering, University of Akron, Akron, Ohio, USA

Highlights  R404A spray cooling is studied using a straight-tube nozzle with expansion chamber.  The substitution of R134a with R404A increases the maximum heat flux by 19%.  The introduction of expansion chamber enhances the maximum heat flux by 18%.  R404A spray cooling exhibits intense temporal and radial non-uniformity.  A region of uniform cooling with a radius of 2 mm appears around the spray center. Abstract Cryogen spray cooling is applied to protect epidermis from thermal damage in laser dermatology. However, R134a shows insufficient cooling capacity for minimizing the laser energy absorption by melanin in darkly pigmented human skin. By contrast, the cooling capacity of R404A can be improved with a low boiling point. This study examined the temporal and spatial variations in surface heat transfer during R404A spray cooling using a straight-tube nozzle with an expansion chamber. Substitution of R134a with R404A increases the maximum heat flux by 19%, whereas introducing an expansion chamber enhances the maximum heat flux by 18%. Results indicate that surface heat transfer during R404A spraying exhibits intense temporal and spatial non-uniformity. A sub-region of uniform cooling with a radius of 2 mm appears around the spray center with a high transient heat flux above 300 kW/m2. This finding can help physicians precisely control the therapy area with enhanced laser energy. Key words：spray cooling; R404A; expansion-chambered nozzle; surface heat transfer; laser dermatology

1.

Introduction Port wine stain (PWS) birthmark is a congenital vascular malformation that occurs in approximately 0.3% of

all newborns (Alper and Holmes, 1983). Based on the principle of selective photothermolysis (Kercher et al., 1983; Kelly and Nelson, 2000), pulsed dye laser with a specific wavelength (typically 585/595 nm) has been the standard treatment for this vascular skin lesion. However, the absorption of laser energy by melanin at these wavelengths can induce unwanted heating of the epidermis, an outcome that may lead to irreversible thermal damage (Chang and

*

Corresponding author. Tel.: +86 029 8266 7326; fax: +86 029 8266 9033 E-mail address: [email protected] (B. Chen) 1

Page 1 of 16

Nelson, 1999; Karapetian et al., 2003), especially for darkly pigmented people. Nelson et al. (1995) proposed cryogen spray cooling (CSC) to reduce the epidermis temperature epidermis, thus allowing for a laser pulse with high energy. Cryogen is spurted in milliseconds on the skin surface, through which superficial layers of the skin (epidermis) can be selectively cooled to minimize or eliminate laser-induced thermal injury (Chang et al., 1998). Non-toxic and environment-friendly cryogen R134a, which has a −26.1 C boiling point at 1 atm, is currently the only commercial cryogen that assists in laser therapy. For darkly pigmented skin, the cooling capacity of R134a is insufficient because of its relatively high boiling point. Clinical studies have shown that only less than 20% of patients can achieve complete PWS clearance (Chen et al., 2012). The treatment failure is attributed to the incomplete understanding of the mechanism of interaction between light and tissue. Another factor is the insufficient cooling from cryogen spray, which is especially significant for Asian people because of the strong laser absorption by darkly pigmented skins. Aguilar et al. (2005) proposed the hypobaric pressure method on the skin surface to improve the cryogen cooling capacity. Our recent result (Zhou et al., 2012) indicated that R134a spray cooling can provide 2.6 times the maximum surface heat flux with a pressure of 0.1 kPa in comparison with the atmospheric pressure at a spray distance of 10 mm. Nevertheless, the implementation of a local vacuum on the epidermis still needs to be developed for clinical applications. Dai et al. (2006) suggested that R404A with a low boiling point (−46.5 C at 1 atm) and high volatility is a possible candidate to enhance the cooling efficiency of CSC in clinical application. Zhou et al. (2012) measured the time-varying R404A spray and observed steady-state spray patterns. The maximum surface heat flux qmax can achieve 400 kW/m2, a result that confirms its better cooling capacity compared with that of R134a (with a qmax of approximately 300 kW/m2). The spray is in a jet-like pattern near the nozzle exit with a high droplet concentration. The effective spray cooling radius of R404A at 50 mm distance is approximately 3 mm, which demonstrates that the R404A spray has better spatial selectivity with a smaller cooling surface than does the R134a spray (corresponding cooling radius = 8 mm). However, few studies are devoted to the transient spray pattern and the radial distribution of surface heat transfer for R404A spray, both of which are closely related to clinical practice. Optimizing the nozzle design can also increase the heat extraction from skin. The current nozzles used in clinical practice are all straight-tube types, with a nozzle diameter of 0.5–1.4 mm. Sher et al. (1977) suggested that an expansion chamber can result in efficient atomization at the nozzle exit by enabling a massive flashing process and sufficient time for bubbles to grow. Bar-Kohany et al. (2004) designed expansion nozzles for diesel spray based 2

Page 2 of 16

on Sher’s theory and explored the optimal expansion chamber volume to achieve the highest superheat degree for flash boiling spray. Compared with a straight-tube nozzle, the well-designed nozzle requires lower spray pressure to achieve a comparative atomization effect. Ersoy and Sag (2014) used an ejector as an expansion device instead of an expansion valve in the R134a refrigeration system and improved the coefficient of performance by 6.2–14.5%. Zhou et al. (2014) visualized the internal flow and spray pattern of an expansion-chambered nozzle with different aspect ratios. The atomization effect becomes optimal for CSC when the aspect ratio of the length to the diameter of the chamber ranges from 1:2 to 2:1. Liquid cryogen deposition and heat extraction from epidermis during CSC are temporally and spatially dependent. Research on CSC has a clinical importance to controlling the heat extraction and the cooling area precisely. Temporal and spatial distributions for the surface heat transfer characteristics of R134a spray are widely investigated to achieve optimal cooling efficiency. Aguilar et al. (2001) measured the steady-state lateral variation in surface temperature for CSC and found a small temperature variation zone with a 6 mm radius at the position of 50 mm from the nozzle tip. This finding was also confirmed in our previous work (Zhou et al., 2012) with a similar effective cooling area of approximately 6 mm radius at the same spray distance. Franco et al. (2005, 2007) systematically studied the spatial and temporal non-uniformity in surface heat transfer caused by the uneven deposition and spread of liquid cryogen. Franco reported a sub-region of uniform heat transfer with a 2 mm radius and a safe epidermal cooling protection zone with a 10 mm diameter for a 30 mm distance and a 60 ms duration spurt. The present work aims to investigate the temporal and spatial variations in surface temperature and heat flux during R404A spray cooling from a nozzle with an expansion chamber that has an aspect ratio of 1:1. The objective is to enhance cooling efficiency in comparison with R134a and a straight-tube nozzle. The temporal and spatial dependent profiles for surface heat transfer are of particular interest to guarantee sufficient cooling protection for the epidermis and help understand the mechanism of heat transfer in spray cooling. This study may have instructional significance for implementing R404A and nozzle design during CSC for laser dermatology surgery.

2.

Experimental methods The experimental system consists of a spray system and measurement facilities. Cryogen R404A and a

specially designed nozzle with an expansion chamber are used in the spray system. A high-speed video camera and a thin-film type-T thermocouple (TFTC) are used in the measurement. The surface heat fluxes are calculated by 3

Page 3 of 16

Duhamel’s theorem.

2.1. Spray system Fig. 1 shows the schematic of the experimental facility for CSC. The spray system consists of a pressure vessel for cryogen storage, a three-dimensional translational electric positioner (WN105TA300M by Beijing Winner Optics Instruments Corp., China) with a space resolution of 8 m, a solenoid electric valve (B2021SBTTO24DVC by Gems, USA) with a response time less than 5 ms, and a specially designed nozzle. The pressure in the cryogen container is set at 1.45 MPa, which is higher than the saturation pressure of R404A at ambient temperature, to prevent the sub-cooled boiling of R404A in the connecting hose. Fig. 2 shows the structure of the nozzle with an expansion chamber. The design of the expansion chamber is illuminated by the research of Sher et al. (1977) and Bar-Kohany et al. (2004); this design is proven to result in an efficient atomization effect and a high superheat degree of spray droplets. Our previous work (Zhou et al., 2015) also confirms that the introduction of an expansion chamber leads to high surface heat flux. The nozzle is made of quartz glass and divided into three parts, namely, inlet tube, expansion chamber, and outlet tube. The geometry of the outlet tube resembles that of commercial nozzles used for CSC in conjunction with dermatological laser treatments, with a length of 60 mm and an inner diameter of 1 mm. The expansion chamber is 10 mm  10 mm in length and width. This dimension can help in the efficient atomization of cryogen and in releasing superheat inside the nozzle to generate a small dispersion outside the nozzle, thereby leading to cold liquid and strong heat transfer on the cooling surface. The spray angle is selected as 90, i.e., the nozzle is installed perpendicularly to the skin phantom because the spray angle affects heat transfer in the range of 15°–90 (Aguilar et al., 2004). According to the numerical simulation of surface heat transfer by Li et al. (2014), an optimal spurt duration of 75–100 ms is recommended to achieve the best cooling effect. The remaining cryogen in the expansion chamber after valve closing provides an additional spurt duration. Thus, the valve opening time should be set shorter than the recommended duration. A trial of 50 ms is performed in this experiment to demonstrate the suitability in the section of spray pattern.

2.2. Camera system A high-speed video camera (HG-100, Redlake MASD Inc., USA) is used to take snapshots of the spray pattern. A PLS-SXE300 Xe lamp with an output light power of 50 W provides illumination for the high-speed camera. The camera and lamp are positioned in the same horizontal plane with a small angle less than 30. The angle is adjusted 4

Page 4 of 16

to ensure that the camera reads the reflection light and captures the clear spray configuration. The camera is placed 1000 mm from the spray axis and views down to capture the entire spray and substrate surface. All snapshots are taken at a speed of 1000 fps and a pixel resolution of 1504 × 1128. The snapshots are captured before valve opening. Spray starting time is defined as the moment when the spray first appears outside the nozzle.

2.3. Temperature measurements and heat flux calculations An epoxy resin substrate (50 mm × 50 mm × 5 mm) is used to substitute for the skin phantom because of its similar thermal properties to human skin (Franco et al, 2005). As shown in Fig. 1, a 2 m-thick TFTC is deposited directly on the epoxy substrate surface to measure the transient surface temperature during the pulsed spray. TFTC satisfies the requirements of high measurement resolution and relatively wide measurement range with a sampling frequency of 100 kHz. The dynamic response of TFTC is tested by using a single Nd: YAG laser pulse with a 6 ns duration. The response increases from ambient temperature to approximately 550 C within 5 s, with a temperature change rate over 100 C/s, which is fast enough to measure the rapid change in surface temperature. The surface temperature of the spray center is measured for different spray distances from 10 mm to 50 mm with an axial interval of 10 mm. For radial distribution research, the measuring points are arranged every 2 mm from the spray center to the periphery of the sprayed surface (Fig. 1). The temperature at each location is measured at least three times to ensure repeatability of the experiment. Data are accepted only when the deviation in each measuring temperature is below 3 °C at a specified time (25, 50, 100, and 150 ms). After acquiring the surface temperature, Duhamel’s theorem is used to evaluate the temperature gradient on the substrate surface and generate the surface heat flux, with the assumption of one-dimensional heat transfer problem along the depth direction. Taler (1996) provided a detailed description for solving the heat flux q, i.e., q (t ) 

 c 



t

0

1 t

dT    d

d

(1)

where T(Θ) is the history of the measured surface temperature. Temperature measurements are conducted at discrete instants (t1, t2, …, tM). Assuming that the surface temperature between successive instants varies linearly with time, Eq. (1) can be integrated analytically to give q (t M 1 )  2

 c 

M



i 1

Ti  1  Ti t M 1  ti 1 

t M 1  ti

(2)

2.4. Uncertainty analysis The thermoelectric property of TFTCs agrees well with that of standard T-type thermocouples. Careful 5

Page 5 of 16

calibration is made by measuring the same location in a thermostatic bath with TFTC and a standard T-type thermocouple, as shown in Table 1. The variation between TFTC and the standard thermocouple is within 0.5%, and the uncertainty of the surface temperature measured by TFTC is within 1.5 °C. The uncertainty of the calculated surface heat flux includes two parts, namely, the stochastic error caused by the uncertainty of temperature measurement and the systematic error caused by the discretization of the integral. Stochastic error Δqsto can be expressed as M 1

 q sto ( t M  1 ) 

 q

  Ti     Ti 

  i 1

(3)

where ΔTi is the uncertainty of temperature measurement, and all ΔTi are assumed to be equal. Substituting Eq. (2) into Eq. (3), we can simplify Δqsto as  q sto  4  T

 c

(4)

 t

where Δt is the time interval. The systematic error Δqsys is caused by the discretization of the integral into a linear equation. Hence, its value equals the difference between Eqs. (1) and (2). Δqsys approaches zero only when function T(Θ) is linear, but the exact value cannot be calculated because the temperature function cannot be obtained. The selection of Δt greatly influences the uncertainty of heat flux. A small Δt diminishes Δqsys because of the good linearity at each time step, while Δqsto is inversely proportional to

t

. Δt is set as 2 ms to restrict the value

of Δqsys and Δqsto to a minimum. Δqsys approximates zero, and Δqsto can be determined by Eq. (4). Therefore, the uncertainty of the calculated surface heat flux is within  5.1 kW/m2.

3.

Results and discussion

3.1. Spray pattern Snapshots of the R404A spray and droplet impacting the substrate surface are shown in Fig. 3. The spray gradually expands within the valve opening time of 50 ms. After the solenoid valve closes, the remaining liquid cryogen inside the nozzle allows the spray outside the nozzle continue until the spray gradually comes to an end at approximately 300 ms, which leaves the liquid film still on the substrate surface. The spray pattern and the covered surface vary intensely with time, i.e., spray characteristics have strong temporal and spatial non-uniformity. The spray pattern is investigated in detail by presenting the time variation in the spray angle and radius. Spray 6

Page 6 of 16

angle α is defined as the angle between two lines connecting the nozzle exit center and the periphery of the spray at 15 mm out from the nozzle (Fig. 4), in reference to the work of Juslin et al. (1995). Spray radius rs is defined as the distance between the spray center and the periphery of the spray on the substrate surface. Fig. 5 shows the variations in α and rs as a function of time with similar tendencies, both achieving a single peak at 50 ms. The peak value of R404A spray angle from the nozzle with an expansion chamber is 53.9, and the peak value of rs is 8.0 mm. The curves start to decline rapidly after the solenoid electric valve closes, while rs continues to increase over 4 mm within 75 ms. The snapshots show that the R404A spray becomes rather dilute after 100 ms. An approximate efficient spurt duration of 75–100 ms can be concluded, which coincides with the predicted time of optimal spray duration by Li’s multiscale model (2014).

For comparison, α and rs of the R134a spray from the same nozzle and the R404A spray from the traditional straight-tube nozzle are also measured. The R404A spray angle from the straight-tube nozzle (without an expansion chamber) can achieve 108.9 and rs can reach 17.3 mm, both of which are higher than those from the nozzle with an expansion chamber. The sub-cooled liquid cryogen intensely evaporates before spurt within the expansion chamber, and thus large superheat is reduced at the nozzle exit. The insignificant evaporation at the nozzle exit leads to a small spray angle and a good cooling effect. In the same nozzle with an expansion chamber, α and rs peak of R134a spray are 77.3 and 12.3 mm respectively. These values are higher than those of the R404A spray and indicate that R404A spray is more concentrated with better spatial selectivity than R134a.

3.2. Axial distribution of surface heat transfer The temperature data obtained by our thin-film thermocouple is given in Fig. 6. The figure shows the time evolution of the surface temperature T of the substrate at the R404A spray center for different spray distances with an axial interval of 10 mm. As shown in the figure, four typical periods of temperature changes exist in CSC for all curves. The delay period lasts less than 10 ms for all distances, when T remains almost constant before droplets reach the substrate surface. This delay includes the response time of the solenoid electric valve, cryogen flow inside the nozzle, and droplet spurting. A rapid dropping period follows, with approximately 5 ms duration, when T falls directly below −30 °C from ambient temperature for all distances. The third period is the cooling maintenance, when T keeps a low temperature with rather gradual variation for a relatively long period. Each curve is remarkably different in this period, with different minimum temperatures and durations. The minimum surface temperature 7

Page 7 of 16

keeps dropping as the axial distance increases (−42.66 °C at z = 10 mm and −57.37 °C at z = 50 mm). The temperature of the R404A droplet is already close to the boiling temperature Tboiling at the nozzle exit. Flash evaporation leads to intense heat absorption from the droplet itself and a continuous decrease in droplet temperature as droplets fly forward (Zhou et al., 2012). The impacting droplet with a low temperature evaporates, spills, or forms a liquid film on the surface and extracts heat from the surface through boiling, convective, or conductive heat transfer. The minimum surface temperature can reach below Tboiling for 20–50 mm. The last period is the slow recovery, when T gradually increases to ambient temperature. The curves for all distances start to rebound at 200–300 ms, which is in accordance with the spray ending time in the snapshots of the spray pattern.

Correspondingly, Fig. 7 shows the time evolution of the calculated surface heat flux q at various spray distances (i.e., z = 10, 20, 30, 40, 50 mm). Similar to surface temperature, q varies with time and space intensively. In all cases, q dramatically reaches the peak qmax during the rapid temperature-dropping period, when droplets violently impact the surface with a high velocity. Afterwards, q decreases at a relatively slower velocity as the liquid film forms and during the heat convection because of the decrease in impacting droplets. The heat flux then begins to fluctuate as the heat transfer process changes from the transition boiling regime into the nucleate boiling regime (Aguilar et al., 2003). q finally drops to zero when the spray ends and the liquid film evaporates completely. Although all curves have a similar varying tendency, differences obviously exist, especially in the first tens of milliseconds. The distance z = 30 mm is selected to be the optimal spray distance, in which the maximum value qmax = 483.04 kW/m2 is acquired. At closer distance, a higher droplet temperature leads to less surface heat extraction. At farther distance, the spray has a larger dispersion range and a more dilute density at the center, and thus the surface heat flux is lower.

A similar varying tendency of surface heat transfer at different spray distances is observed in Fig. 7. A non-dimensional correlation is expected to calculate the surface heat transfer at different distances. Fig. 8 shows the relation of Nuc/Nuc,max as a function of normalized time τ, where subscripts c and max represent the value at the spray center and the maximum value respectively. The local Nusselt number and the normalized time are respectively defined as Nu 

q

rs

T  Ta 

,

 

t  t0 t m ax  t 0

(5) 8

Page 8 of 16

where Ta is the ambient temperature, tmax is the time when Nuc,max occurs, and t0 is the initial time when Nuc first exceeds 5% of Nuc,max. The introduction of t0 eliminates the different delay times at each spray distance and obtains similar curves. The piece-wise representation of the curves can be summarized as follows:

Nuc N u c , m ax

  1  0.44    1     0.56  0.08    2    0.32  0.044    5   0.2  0.006   10   

 1 1<   2 2<   5

(6)

5<   10

  10

The surface heat fluxes among different cryogens and nozzles are compared in Fig. 9, and all data are measured at the spray center at z = 30 mm. With the same straight-tube nozzle, R404A has a longer increasing time for q and a higher peak value than R134a. With the same cryogen R404A, the nozzle with an expansion chamber has a higher peak value than the straight-tube nozzle. The substitution of cryogen increases qmax by 19%, and the improvement of the nozzle increases qmax by 18%. The heat transfer enhancement by substituting cryogen is the joint effect of the thermal properties of cryogen, the fraction and velocity of the mixture gas, and the thickness of the liquid film, all of which change the evaporative, convective, and conductive surface heat transfers. The R404A spray has smaller droplet density and diameter and higher droplet velocity than the R134a spray (Zhou et al., 2013). This finding indicates a negative influence in forming the liquid film and a positive effect in enhancing the evaporative and convective surface heat transfers. The enhancement mechanism of the expansion chamber is analyzed according to a high-speed video for bubbles within the expansion chamber in Zhou et al. (2015). A violent phase change and large bubbles can be observed within the expansion chamber, which means that the liquid forms the first flashing breakup within the expansion chamber and the second flashing spray outside the nozzle exit. Therefore, the expansion-chambered nozzles can cause lower droplet temperature at the nozzle exit, lower surface temperature, and higher heat flux compared with the straight-tube nozzle.

3.3. Radial distribution of heat transfer Fig. 10 displays the radial and temporal distributions of surface temperature T. The axial distance is fixed at 30 mm, which is the optimal spray distance to acquire the best cooling efficiency. As shown in the figure, the temperature difference between the spray center and the radial location r = 2 mm is consistently lower than 3 °C. When far from the spray center, the temperature usually obtains a higher value. The location r = 10 mm maintains a 9

Page 9 of 16

temperature higher than 10 °C, which can be considered out of the spray area. A sudden temperature change occurs at r = 8 mm, which is exactly the peak value of rs and can be confirmed as the periphery of the cooling area. The same four temperature changing periods appear at all radii, except at r = 10 mm, which leads to a delay at the first 10 ms, a rapid drop within 5 ms, cooling maintenance, and a slow recovery during several hundreds of milliseconds. The surgery time for PWS is usually less than 150 ms, a time when a low temperature value (less than −30 °C) can be maintained near the spray center (r ≤ 4 mm).

Correspondingly, the radial and temporal distributions of surface heat flux q are drawn in Fig. 11. Same as surface temperature, q acquires a nearly equal value for the area r ≤ 2 mm. Hence, a sub-region with a radius of 2 mm appears on the cooling surface in the R404A spray, where the heat transfer characteristics are uniform. However, q varies intensively with the radial distance out of the spray center, and the maximum difference for different radial locations can achieve over 400 kW/m2. High heat flux only appears at a small radius and within tens of milliseconds. Most of the heat extracted from the surface occurs during the first 50 ms of the spray. Surface cooling of the R404A spray is a transient small-scale process with a time scale of 10−3 s and a length scale of 10−3 m. Combining the spray patterns presented in Figs. 3 and 5 illustrates that the radial and temporal non-uniformity characteristics of surface heat transfer are closely related to the spray pattern. At the first 25 ms, α and rs keep less than 30 and 4 mm respectively; the small values lead to the concentration of R404A at the center spray area. Therefore, low temperature (less than −30 °C) and high heat flux (larger than 250 kW/m2) occur only in the area with a radius of 4 mm. At 25–100 ms, high values of α and rs result in a small cryogen concentration and a wide cooling region. As a result, the spray area with a radius of 6 mm maintains a low temperature but with low heat flux. The cooling region then gradually reduces while keeping the same pace as rs.

The radial distributions of surface temperature and heat flux at four instants are introduced here, as shown in Fig. 12. The surface temperature reaches minimum around the spray center and almost increases monotonously with radial distance at each instant. In Fig. 12(a), the temperature distributions obey the Gaussian distribution, which is also described in literature (Choi, 2002). At the valve closing instant (50 ms), the temperature curve is entirely lower than that at 25 ms, while each location does not acquire the lowest temperature. In the therapy duration (~150 ms), the temperature at the spray center continues to drop, while the temperature around the external 10

Page 10 of 16

edge of the spray gradually increases after the valve closes. The surface temperature within a 6 mm radial distance is lower than −20 °C at 150 ms, which is still suitable for clinical therapy. All curves are almost parallel to the axis around the spray center, which gives the radius scale of the sub-region a uniform surface temperature. In Fig. 12(b), the surface heat flux almost decreases monotonously with radial distance, and the curve continues falling overall with increasing time as the heat flux achieves the peak value before 25 ms. All the curves are also parallel to the axis around the spray center, an outcome that indicates the surface heat transfer is uniform in area with a radius of 2 mm from the spray center at all times and 4 mm from the spray center in the duration of valve opening (50 ms). This experimental result is similar to that in Franco’s research (2005) on R134a spray cooling with a straight-tube nozzle. More detailed surface heat transfer characteristics in R404A spray cooling are presented in our previous work (Wang et al., 2015). Similar to the dimensionless treatment for heat transfer coefficient at the spray center at different distances, normalized Nur/Nuc,max as a function of normalized time τ at different radial locations can also be drawn, where subscript r represents the radial location. In Fig. 13, the five curves correspond to normalized Nur/Nuc,max at locations from the spray center to the external edge with a radial interval of 2 mm. Two phases are displayed in these curves, namely, climbing approximately linearly to the peak and declining with an approximately equal rate. This process can be easily described as

Nur N u c , m ax

  N uc   Nu  c , m ax r , m ax   N u c   N u c , m ax 

   r , m ax

(7)    r , m ax

where τr,max represents the dimensionless time when Nur achieves its maximum value. The values of τr,max recommended for this particular experimental condition are listed in Table 2.

The comparisons of heat extraction per area Q versus time at different radial locations are presented in Fig. 14. All heat extractions first monotonically increase with time once the droplets impact the substrate surface. The increasing rate monotonically reduces from the spray center to the periphery in this varying phase. As time prolongs, the increasing rate becomes gradual because of the poor heat conductivity of the liquid layer formed on the surface. All heat extractions eventually achieve stable value as the liquid cryogen on the surface completely 11

Page 11 of 16

evaporates. The time to reach such a final value is approximately 300 ms for all locations, except for r = 4 mm (around 400 ms). The last liquid cryogen remains at r = 4 mm for an intense splash of cryogen around the center and a high evaporation rate near the periphery. The final stable Q is rather close for the area r ≤ 4 mm and reduces by half at r = 6, 8 mm. Heat extraction per area at the spray center is approximately 30 kJ/m2, which will not cause cold injury to the epidermis.

4.

Conclusions Compared with R134a, R404A can improve the cooling protection to the epidermis because of its lower

boiling point and smaller spray dispersion. Research on surface heat transfer characteristics during R404A spray has an instructional significance in improving the laser treatment clinical efficacy for people with darkly pigmented skin. Experiments are conducted to obtain the spray pattern and surface temperature during R404A spray cooling, as well as investigate the temporal and radial profiles for surface heat transfer. A straight-tube nozzle with an expansion chamber is introduced to CSC to improve the cooling efficiency. The conclusions are as follows: 1) The spray of R404A is more concentrated than that of R134a, and the cryogen spray from the nozzle with an expansion chamber is more concentrated than that from the straight-tube nozzle. The application of R404A spray from the nozzle with an expansion chamber can provide better skin surface cooling efficiency in laser surgery. The substitution of cryogen increases qmax by 19%, and the improvement of the nozzle increases qmax by 18%. 2) Heat transfer on the cooling surface during the R404A spray exhibits an intense spatial and temporal non-uniformity. The optimal spray distance to acquire the best cooling efficiency is confirmed to be 30 mm, with which the maximum heat flux can achieve 483.04 kW/m2. The maximum radial temperature difference is nearly 60 °C, and the heat flux difference can achieve over 400 kW/m2. The maximum surface heat flux appears at tens of milliseconds, and most of the heat extracted from the skin occurs during the first 50 ms of the spray. The surface maintains a low temperature within 150 ms (therapy time) because of the liquid film spread out on the surface and the long-time spray duration owing to the expansion chamber. 3) A similar varying tendency of surface heat transfer at different spray distances and radial locations is generated. The correlations of normalized surface heat transfer coefficient for R404A spray are summarized, which are meaningful for our future simulation work. 12

Page 12 of 16

4) The spray cooling of R404A is a transient small-scale process with a time scale of 10−3 s and a length scale of 10−3 m. Accumulated heat extraction in the central region with a radius of 4 mm is uniform, with a per area value of approximately 30 kJ/m2, which will not cause cold injury to the epidermis. A sub-region of uniform cooling effect with a radius of 2 mm appears around the spray center, an effect that can help physicians precisely control the therapy area with enhanced laser energy.

Acknowledgement This work was supported by the National Natural Science Foundation of China (51336006), the International Science & Technology Cooperation Plan of Shaanxi Province (2013KW30-05) and Fundamental Research Funds for the Central Universities.

References Alper, J.C., Holmes, L.B., 1983. The incidence and significance of birthmarks in a cohort of 4,641 newborns. Pediat. Dermatol. 1, 58-68. Aguilar, G., Verkruysse, W., Majaron, B., Svaasand, L.O., Lavernia, E.J., Nelson, J.S., 2001. Measurement of heat flux and heat transfer coefficient during continuous cryogen spray cooling for laser dermatologic surgery. IEEE J. Sel. Top. Quantum Electr. 7, 1013-1021. Aguilar, G., Wang, G.X., Nelson, J.S., 2003. Effect of spurt duration on the heat transfer dynamics during cryogen spray cooling. Phys. Med. Biol. 48, 2168-2181. Aguilar, G., Vu, H., Nelson, J.S., 2004. Influence of angle between the nozzle and skin surface on the heat flux and overall heat extraction during cryogen spray cooling. Phys. Med. Biol. 49, N147-N153. Aguilar, G., Svaasand, L.O., Nelson, J.S., 2005. Effects of hypobaric pressure on human skin: feasibility study for port wine stain laser therapy (part I). Laser Surg. Med. 36, 124-129. Aguilar, G., Franco, W., Liu, J., Svaasand, L.O., Nelson, J.S., 2005. Effects of hypobaric pressure on human skin: implications for cryogen spray cooling (part II). Laser Surg. Med. 36, 130-135. Bar-Kohany, T., Sher, E., 2004. Subsonic effervescent atomization: A theoretical approach. Atomization Sprays 14. Chang, C.J., Anvari, B., Nelson, S.J., 1998. Cryogen spray cooling for spatially selective photocoagulation of hemangiomas: a new methodology with preliminary clinical reports. Plast. Reconstr. Surg. 102, 459-463. Chang, C.J., Nelson, J.S., 1999. Cryogen spray cooling and higher fluence pulsed dye laser treatment improve port-wine stain clearance while minimizing epidermal damage. Dermatol. Surg. 25, 767-772. 13

Page 13 of 16

Chen, J.K., Ghasri, P., Aguilar G., Drooge, A.M., Wolkerstorfer, A., Kelly, K.M., Heger, M., 2012. An overview of clinical and experimental treatment modalities for port wine stains. J. Amer. Acad. Dermatol. 67, 289-304. Choi, B., Welch, A.J., 2002. Infrared imaging of 2-D temperature distribution during cryogen spray cooling. J. Biomech. Eng. 124, 669-675. de Souza, A.G.U., Barbosa, J.R., 2013. Experimental evaluation of spray cooling of R-134a on plain and enhanced surfaces. Int. J. Refrigeration 36, 527-533. Dai, T., Yaseen, M.A., Diagaradjane, P., Chang, D.W., Anvari, B., 2006. Comparative study of cryogen spray cooling with R-134a and R-404a: implications for laser treatment of dark human skin. J. Biomed. Opt. 11, 041116-041127. Ersoy, H.K., Sag, N.B., 2014. Preliminary experimental results on the R134a refrigeration system using a two-phase ejector as an expander. Int. J. Refrigeration 43, 97-110. Franco, W., Liu, J., Wang, G.X., Nelson, J.S., Aguilar, G., 2005. Radial and temporal variations in surface heat transfer during cryogen spray cooling. Phys. Med. Biol. 50, 387-397. Franco, W., Liu, J., Romero-Méndez, R., Jia, W., Nelson, J.S., Aguilar, G., 2007. Extent of lateral epidermal protection afforded by a cryogen spray against laser irradiation. Laser Surg. Med. 39, 414-421. Juslin, L., Antikainen, O., Merkku, P., Yliruusi, J., 1995. Droplet size measurement: I. Effect of three independent variables on droplet size distribution and spray angle from a pneumatic nozzle. Int. J. Pharm. 123: 247-256. Kercher, D.M., Sheer, R.E., So R.M.C., 1983. Short duration heat transfer studies at high free-stream temperatures. J. Eng. Gas Turb. Power-T. ASME 105, 156-166. Kelly, K.M., Nelson, J.S., 2000. Update on the clinical management of port wine stains. Laser Surg. Med. 15, 220-226. Karapetian, E., Aguilar, G., Kimel, S., Lavernia, E.J., Nelson, J.S., 2003. Effects of mass flow rate and droplet velocity on surface heat flux during cryogen spray cooling. Phys. Med. Biol. 48, N1-N6. Li, D., Chen, B., Wu, W.J., Wang, G.X., He, Y.L., 2014. Multi-scale modeling of tissue freezing during cryogen spray cooling with R134a, R407c and R404A. Appl. Therm. Eng. 73, 1489-1500. Nelson, J.S., Milner, T.E., Anvari, B., Tanenbaum, B.S., Kimel, S., Svaasand, L.O., Jacques, S.L., 1995. Dynamic epidermal cooling during pulsed laser treatment of port-wine stain: A new methodology with preliminary clinical evaluation. Arch. Dermatol. 131, 695-700. Taler, J., 1996. Theory of transient experimental techniques for surface heat transfer. Int. J Heat Mass Transfer 39, 3733-3748. Sher, E., Elata, C., 1977. Spray formation from pressure cans by flashing. Ind. Eng. Chem. Proc. Design Dev. 16, 237-242. Wang R., Zhou Z.F., Bai F.L., Chen B., Wang G.X., 2015. Temporal and radial dependent profiles for surface heat transfer during R404a spray cooling. J. Chem. Ind. Eng. 66, 1258-1264. Zhou, Z.F., Wu, W.T., Chen, B., Wang, G.X., Guo, L.J., 2012. An Experimental Study on the Spray and Thermal Characteristics of R134a Two-Phase Flashing Spray. Int. J Heat Mass Transfer, 55, 4460-4468. Zhou, Z.F., Wang, G.X., Chen, B., Wang Y.S., Guo L.J., 2012. Effect of Pressure and Distance on the Dynamics of Spray Formation 14

Page 14 of 16

and Heat Transfer of R134a Pulsed Flashing Spray. 7th International Symposium on Multiphase Flow, Heat Mass Transfer and Energy Conversion, Xi’an, China. Zhou, Z.F, Chen, B., Wang, Y.S., Guo, L.J., Wang, G.X., 2012. An experimental study on pulsed spray cooling with refrigerant R-404a in laser surgery. Appl. Therm. Eng. 39, 29-36. Zhou Z.F., Wang R., Chen B., Wang G.X., Guo L.J., 2013. Comparative Study on Spray Characteristics and Heat Transfer Dynamics of pulsed Spray Cooling with Volatile Cryogens. International Workshop on Heat Transfer Advances for Energy Conservation and Pollution Control, Xi’an, China. Zhou, Z.F., Bai, F.L., Wang, R., Chao, R., Chen, B., Wang, G.X., 2014. Visualization of the Flashing Spray Generating by the Expansion-chamber Nozzle using R134a. Annual Conference of Chinese Society of Engineering Thermophysics, Xi’an, China. Zhou, Z.F, Chen, B., Bai, F.L., Wang, R., Wang, G.X., 2015. Heat transfer dynamics of R404a flashing pulsed spray cooling using expansion-chamber nozzle. J. Chem. Ind. Eng. 66, 100-105. Fig.1 - Schematic of experiment system

Fig.2 - Nozzle structure Fig.3 - Snapshots of R404A spray pattern Fig.4 - Definition of spray angle α and spray radius rs

Fig.5 - Variation of spray angle α and spray radius rs Fig.6 – Surface temperature at the spray center for different spray distances Fig.7 – Surface heat flux at the spray center for different spray distances Fig.8 – Normalized heat transfer coefficient Nuc/Nuc,max as a function of normalized time τ for different spray distances and the corresponding piece-wise representation Fig.9 - Surface heat flux at the spray center Fig.10 – Radial and temporal distributions of surface temperature Fig.11 – Radial and temporal distributions of surface heat flux Fig.12 - Radial distributions for surface temperature and heat flux Fig.13 – Normalized heat transfer coefficient Nur/Nuc,max as a function of normalized time τ at different radial locations Fig.14 – Time variations of heat extraction per area Q at different radial locations Table 1 - Comparison of temperature measured by TFTC and standard thermocouple Standard thermocouple/K

TFTC/K

Relative deviation/%

221.20

220.34

-0.39

15

Page 15 of 16

237.50

236.67

-0.35

251.78

250.71

-0.42

266.09

266.39

0.11

283.70

282.42

-0.45

300.54

299.83

-0.24

312.91

313.63

0.23

Table 2 - τr,max for different radial locations r (mm)

0

2

4

6

8

τr,max

1.0

1.1

1.5

3.0

4.5

16

Page 16 of 16