International Journal of Thermal Sciences 110 (2016) 146e158

Contents lists available at ScienceDirect

International Journal of Thermal Sciences

journal homepage: www.elsevier.com/locate/ijts

Analytical solutions of non-Fourier bio-heat conductions for skin

subjected to pulsed laser heating
Shueei-Muh Lin a, *, Chang-Yu Li b
a
KunShan University, Tainan, 710-03, Taiwan, Republic of China
b
Guangzhou College of South China University of Technology, Peoples Republic of China

a r t i c l e i n f o a b s t r a c t

Article history: The analytical solutions of bio-heat transfer for skin tissue with general boundary conditions in the
Received 7 October 2015 Pennes, Cattaneo-Vernotte (C-V) and Dual-phase-lag (DPL) models are presented. The heat transfer of
Received in revised form skin subjected to the pulse laser heating and uid cooling are studied. In the conventional literature, the
24 June 2016
inner boundary of skin will be assumed to be in the constant temperature condition or the isolated
Accepted 28 June 2016
Available online 14 July 2016
condition and their numerical difference is discussed here. The effects of the phase lags with respect to
heat ux and temperature gradient on the thermal wave p speed are investigated. The comparison of the

presented results and the approximated speed formula, k=tq rc, in the C-V model is made. The skin
Keywords:
Bio-heat transfer
damage index in several damage models by pulsed laser heating is discussed. Later, the discussion of
Dual-phase-lag model effects of several boundary conditions on temperature variation and thermal damage in the Pennes, CV
Thermal wave speed and DPL models is introduced. It is found that the procedure including pulsed heating and cooling creates
Analytical solution the temperature uctuation of skin. It is suitable for the cold and heat therapy for pain relief and other
Thermal damage functions. The conditions of pulsed heating skin and simultaneous cooling for the thermal therapy are
Pulse laser heating investigated. The analytical method can be widely applied for solving the general problem of heat
Thermal therapy conduction.
2016 Elsevier Masson SAS. All rights reserved.

1. Introduction Pennes model as it assumes that the propagation speed of thermal

The heat transfer in skin tissue is a very complex process. It
involves multiple phenomenological mechanisms including con-
duction in tissue, convection between blood and tissues, blood
perfusion or advection and diffusion through micro vascular beds,
and metabolic heat generation [1]. The analysis of heat transfer and
thermal damage in skin tissue are of great importance and can be
useful for heat therapy in clinics.
In general, there are three kinds of bio-heat transfer models for
skin tissues:(1) the Pennes model [2], (2) the C-V model [3] and (3)
the DPL model [4]. Firstly, the heat conduction in biological tissue is
modeled in the Pennes model which is based on the Fouriers law
   
! !
q r ; t kVT r ; t
The Pennes model was widely used in bio-heat transfer because
of its simplicity. However, there are some disadvantages for the
* Corresponding author.
E-mail address: smlin45@gmail.com (S.-M. Lin).
disturbance is innite. In fact, heat is always found to propagate
with a nite speed within living biological tissues as they have
highly non-homogeneous inner structure [5].
Secondly, using the concept of a nite heat propagation speed,
Cattaneo [6] and Vernotte [7] simultaneously modied the Fourier
law to form the C-V model:
   
! !
q r ; t tq kVT r ; t (2)
Eq. (2) assumes that the heat ux and the temperature gradient
occur at different times. The delay between the heat ux and
temperature gradient is dened as the thermal relaxation time tq.
Several literature are devoted to determine the phase lag for the
(1)
heat ux [8e12]. The thermal wave theory ensures a strong path
dependency for the temperature gradient. Although the C-V model
can capture the microscale response in time, but not in space [13].
The C-V model is debatable in view of the fast-transient response
with microstructural interaction effects [14].
Thirdly, in order to consider the effect of microstructural in-
teractions in fast transient process of heat transport, Tzou [14]

http://dx.doi.org/10.1016/j.ijthermalsci.2016.06.034
1290-0729/ 2016 Elsevier Masson SAS. All rights reserved.
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158 147

Nomenclature T temperatures of cooling uid

u step function
c specic heat of skin tissue, x position coordinate
cb specic heat of blood x dimensionless position coordinate, x/L
fi dimensionless heating at the boundary, LFi(t)/kTa r density of skin tissue
h heat convective coefcient rb density of blood
k thermal conductivity of skin tissue t dimensionless time,kt/rcL2
L thickness of skin tq phase lag for heat ux, a=Ct2
Nu0, NuL Nusselt number,h0L/k, hLL/k tq dimensionless phase lag for heat ux, ktq/rcL2
p heat transfer per unit area of pulsed laser heating tT phase lag for temperature gradient
qmet metabolic heat generation in the skin tissue tT dimensionless phase lag for temperature gradient, ktT/
qmet dimensionless metabolic heat generation in the skin rcL2
tissue, qmetL2/kTa j ratio of the thermal conductivity to the heat capacity,
qext heat source due to external heating k/rcL2
qext dimensionless heat source due to external heating, Q dimensionless temperature of skin, (T  Ta)/Ta
qextL2/kTa U index of thermal injury,
t time variable 6b the blood perfusion rate
Ta,T temperatures of arterial blood and skin tissue ub dimensionless blood perfusion rate, L2 6b rb cb =k
respectively

established a dual phase thermal lag (DPL) model that allows either
the temperature gradient to precede heat ux vector or the heat
ux vector precede temperature gradient. The model is expressed
as
   
! !
q r ; t tq kVT r ; t tT (3)
where tq is the phase lag for the heat ux vector, and tT is the phase
lag for the temperature gradient. If the local heat ux vector results
in the temperature gradient at the same location but an early time
(tq > tT), the heat transfer is gradient-precedence type. On the
other hand, if the temperature gradient results in the heat ux at a
later time (tq < tT), the heat ow is called ux-precedence type.
Antaki [4] and Xu et al. [14]have proposed dual-phase-lag model of
bio-heat transfer (DPL) that allowed either the heat ux vector
precede temperature gradient or the temperature gradient to pre-
cede heat ux vector. Afrin et al. [1] presented a generalized dual-
phase lag model for living biological tissues based on non-
equilibrium heat transfer between tissue, arterial and venous
bloods. It was found that the phase lag time is greatly dependent to
the properties of tissue and blood.
Due to its complexity, the approximated methods such as the
meshless radial basis collocation method (RBCM), the nite
element method (FEM), the nite difference method (FDM), the
boundary element method (BEM) and the statistical method are
often considered in solving the complex bio-heat transfer 3D
problems. However, the analytical solution is helpful for precisely
investigating the problem. The solution methods are reviewed as
follows:
Xu et al. [15] and Ng and Chua [16] investigated the Pennes bio-
heat transfer model by using the nite difference method and the
nite element method. Deng and Liu [17] have obtained analytical
solutions for Pennes bio-heat transfer equation by using Greens
function method. Ozen et al. [18] investigated the bio-heat transfer
in the C-V model by using nite difference method. Ahmadikiaet al.
[19] presented the analytical solution of heat conduction on skin
tissue in C-V model. Xu et al. [20] employed the nite difference
method to solve DPL bio-heat transfer equation. Liu et al. [21]
solved DPL bio-heat transfer equation by using method of Laplace
transform. Lin [22] presented the analytical method for the heat
conduction of skin subjected to harmonic heating in the dual-phase
lag model. Kumar et al. [23] solved DPL bio-heat transfer equation
by Finite element wavelet Galerkin method. Hooshmand et al. [24]
solved the DPL bio-heat transfer with the isolated boundary con-
dition by employing the separation of variables and Duhamels
integral method for both absorbing and scattering tissues. Kumar
et al. [25] solved the Pennes, C-V and DPL models for the heat
conduction of skin with three different constant boundary condi-
tions by the Laplace transform method. The boundary condition is
not time-dependent. The values of metabolic and spatial heat
source in boundary condition of different thermal therapies were
evaluated. Dombrovsky and Timchenko [26] investigated the
problem of laser induced hyperthermia of supercial tumors. So far,
no analytical method is presented for the DPL heat conduction of
skin subjected to time-dependent pulsed laser heating along with
general boundary conditions.
In this paper, the analytical solutions of the Pennes, C-V and DPL
models for pulsed laser heating will be derived. The inuence of
boundary conditions and phase lags on the prediction of temper-
ature and thermal damage will be investigated. The conditions of
pulsed laser heating skin for the thermal therapy will also be
investigated.
2. Dual-phase-lag model
Governing equation in the DPL model is [14,22]
v2 T v2 Tx; t v3 Tx; t
rctq k tT k  6b rb cb T  Ta
vt 2 vx 2 vtvx2

  vT vqmet
 tq 6b rb cb rc qmet qext tq
vt vt

vqext
tq ; 0 < x < L; t > 0: (4)
vt
where r, c, k are the density, specic heat and thermal conductivity
of skin tissue, respectively; rb, cb are the density and specic heat of
blood, 6b is the blood perfusion rate (volume blood per unit mass of
tissue per unit time); Ta and Tare the temperatures of arterial blood
and skin tissue respectively; qmet is the metabolic heat generation
in the skin tissue and qext is the heat source due to external hea-
ting.tq is the phase-lag in establishing the heat ux and associated
conduction through a medium.tT is the phase-lag in establishing
148 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158

the temperature gradient across the medium. 3. Solution method

The general boundary conditions are
3.1. Change of variable
vT0; t
At x 0 : k h0 T  T F~1 t; (5)
vx To nd the solution for the second-order differential equation
with non-homogeneous boundary conditions, one extends the
vTL; t shifting variable method developed by Lin [21] by taking
At x L : k  hL T  Ta 0; (6)
vx
qx; t vx; t g1 xf 1 t (16)
where the coupled effect of conduction and convection is consid-
ered. At the skin surface, the time-dependent heat source is given. where g1(x) are two shifting functions and v (x, t) is the transformed
The initial condition is function.
Substituting Eq. (16) into Eqs. (8)e(11), two following sub-
vTx; 0 systems are obtained. The rst subsystem is expressed in terms of
Tx; 0 37o C; 0; (7)
vx the transformed variable v(x,t) as follows: the transformed gov-
In terms of the dimensionless parameters listed in the nomen- erning equation is
clature, the dimensionless governing equation is
v2 v v2 v v3 v   vv
v q 2
v q 2
v q 3 
 vq tq tT  ub vx; t  tq ub 1 qts (17a)
tq 2 2 tT  ub q  tq ub 1 vt2 vx2 vtvx
2 vt
vt vx vtvx
2 vt
  (8)
vqmet vqext where
qmet qext tq tq
vt vt  
vq vq
qts qmet qext tq met tq ext
Obviously, if tT 0, the dual-phase-lag model becomes the C-V vt vt
model. Further, if tq 0, the C-V model becomes the Fourier model.
" # (17b)
  df 1 d2 f 1
The dimensionless general boundary conditions are  g1 ub f 1 tq ub 1 tq 2
dt dt
At x 0:

vq0; t and the associated boundary conditions are

g12  g11 q0; t g11 q g12 ~f 1 f 1 (9) At x 0:
vx
Nu0
where g11 1N 1 .
and g12 1N vv0; t
u0 u0 g12  g11 v0; t 0 (18)
At x 1: vx
At x 1:
vq1; t
g22  g21 q1; t 0 (10)
vx
vv1; t
g22  g21 v1; t 0 (19)
where g21 NuL
and g22 1 vx
1NuL 1NuL .
The dimensionless initial conditions are The corresponding initial conditions are

vqx; 0
qx; 0 0; 0; (11) vx; 0 g1 xf 1 0 (20)
vt
Several limiting boundary conditions are described as follows:
Case 1: If g11 1 and g12 0, the general boundary condition (9) vvx; 0 df 0
g1 x 1 (21)
becomes the constant temperature condition, vt dt
The second subsystem is expressed in terms of the shifting
q0; t q (12) function gi(x). The transformed governing equation is
Case 2: If g11 0 and g12 1, the general boundary condition (9)
becomes the heating boundary condition, d2 g1
0; 0<x<1 (22)
dx2
vq0; t ~
f1 (13)
vx and the associated boundary conditions
At x 0:
Case 3: If g21 1 and g22 0, the general boundary condition
(10) becomes the constant temperature condition [14]
dg1 0
g12  g11 g1 0 1; (23)
q1; t 0 (14) dx

Case 4: If g21 0 and g22 1, the general boundary condition (9) At x 1:

becomes the isolated boundary condition [19]
dg1 1
g22  g21 g1 1 0 (24)
vq1; t dx
0 (15)
vx
The shifting function is derived as follows:
The solution of Eq. (22) is assumed
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158 149

g1 x a10 a11 x (25)

g12 l  g22 l sin l g21 cos l g11 g22 l cos l g21 sin l 0
(35)
Substituting Eq. (25) into the boundary conditions (23e24), one
Moreover, the coefcients {C1, C2} are determined and the
obtains
eigenfunction is derived

1 xg21 g22 g11

g1 x (26) Yn x cos ln x sin ln x (36)
g12 g21 g11 g21  g22 ln g12
The orthogonality of eigenfunctions is easily proved and
expressed as
Z1
0; for msn
3.2. Solution of the transformed variable Ym Yn dx ; (37)
dn ; for m n
0
3.2.1. Orthogonalty of eigenfunctions
The characteristic equation of Eq. (17) is Where
! !
2 2
2
v v v v 2 dn 12 g212 g112 g11 g2 12 sin4l2nln g212  g112  g11 g212 cos 2ln .
tq : (27) ln ln ln 2ln
vt2 vx2

By considering the separation of variable method, the solution 3.2.2. Mode superposition method
of Eq. (27) is expressed as Based on the orthogonality conditions (37), the mode super-
position method is used to derive the solution of the transformed
vx; t YxGt (28) system composed of Eqs. (17)e(21). The transformed variable is
assumed to be
Substituting Eq. (28) into Eq. (27), one obtains
X

vx; t Yn xBn t (38)
d2 Y 2 n0
2
l Y 0 (29)
dx By substituting Eq. (38) into Eq. (17) and multiplying it by Ym
and integrating it from 0 to 1, Eq. (17) becomes
d2 G l2
G 0; (30) d2 Bm  2   dBm  2 
dt2 tq tq t T lm tq u b 1 lm u b Bm qm (39)
dt2 dt
Further, substituting Eq. (28) into Eqs. (18) and (19), the corre-
R1
sponding boundary conditions are where qm d1 0 qts Ym dx.
At x 0:
m

Further, the general solution of Eq. (39) is easily derived

dY0 dBm 0
g12  g11 Y0 0 (31) Bm t Vm1 tBm 0 Vm2 t
dx dt
at x 1:
Zt
1
Vm2 t  cqm cdc (40)
tq
dY1 0
g22  g21 Y1 0 (32)
dx
where the fundamental solutions {Vm1, Vm2} of Eq. (39) are deter-
Based on Eqs. (29), (31) and (32), one can derive the eigen- mined as follows:
function as Case 1:If tT l2m tq ub 12  4tq l2m ub > 0, the funda-
mental solutions are
Yx ebx (33)
1
By substituting Eq. (33) into Eq. (29), one obtains Vm1  c21 ec11 t c11 ec21 t ;
c11  c21
(41a)
1
Y C1 cos lx C2 sin lx (34) Vm2  ec11 t ec21 t 
c11  c21
According to the boundary conditions (31e32), the frequency
equation is obtained where

2 r3
2      2  
6 tT lm tq ub 1  tT l2m tq ub 1  4tq lm ub 7
2

c11 6
4
7;
5
2tq

2 r3 (41b)
     2  
6 tT l2m tq ub 1 tT l2m tq ub 1  4tq l2m ub 7
c21 6
4
7:
5
2tq
150 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158

dBm  2 
Case 2:If tT l2m tq ub 12  4tq l2m ub < 0, the funda- lm ub Bm qm ; (44)
dt
mental solutions are
" #

 R1
c where qm d1 qts Ym dx and qts qmet qext  g1 ub f 1 dfdt1 .
Vm1 ec12 t cos c22 t 12 sin c22 t ; m 0
c22
(42a) The solution of Eq. (40) is
1
Vm2 ec12 t sin c22 t
c22 2 8 2
  Zt   <
 lm ub t 4
2
l2m ub t 1 4
where Bm t e Bm 0 e qmet
:dm
   0
tT l2m tq ub 1 ! Z1 39 3
Z1 =
c12 ; 1 df
2tq qext Ym dx  ub f 1 1 g1 Ym dx5 dt5
v
dm dt ;
u  2  32 (42b) 0 0
u l2 u t l 2
t u 1
u m b T m q b
(45)
c22 t 4 5
tq 2tq
where
The dimensionless initial conditions are

21 1 3
g g21 g12 2  g11  g21 g12 g21 g11 cos ln
6 l2 11 ln 7
f1 0 6 n 7
Bm 0 6 " # 7 (43a)
dm g12 g21 g11 6
4 g12 g21 g12 g21 g11
7
5
 2
 3
sin ln
ln ln ln

Z1
g12 g
dBm 0 Yn xdx sin ln  11 cos ln  1;
0 (43b) ln l2n
dt 0

Z1 " ! #
1 1 1 g12 g21 g12 g21 g11
g1 xYm xdx g g21 g12 2  g11  g21 g12 g21 g11 cos ln   sin ln :
g12 g21 g11 l2 11 l ln l2n l3n
0 n n

By substituting Eqs. (26), (38) and (40) into Eq. (16), the
analytical solution of the general system in the dual-phase-lag
model is derived. Moreover, if tT 0, this solution in the dual-
phase-lag model will reduce to that in the C-V model.
5. Evaluation of thermal injury

The burn evaluation is one of the most important characteristics

4. Pennes model in the bioengineering science in skin tissue. The Arrhenius burn
integration proposed by Henriques and Moritz [27] is often used to
Setting tq tT 0, the DPL model will reduce to the Pennes quantify the thermal damage as follows: The dimensionless mea-
model. Eq. (39) becomes sure of thermal damage, U, is the index for burn evaluation
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158 151

Table 1
Values of activation energy and frequency factor for the different injury models.

Injury model Temperature range ( C) Activation energy, Ea (J/kmol) Frequency factor A (1/s)

Henriques [31] All T 6.27  108 3.1  1098

Fugitt [32] T  55 6.27  108 3.1  1098
T > 55 2.96  108 5.0  1045
Stoll and Greene [33] T  55 7.82  108 2.185  10124
T > 55 3.27  108 1.823  1051
Takata [34] T  55 4.18  108 4.322  1064
T > 55 6.69  108 9.389  10104
Wu [35] T  55 6.27  108 3.1  1098
T > 55 6.27  108e5.10  105  (Te53) 3.1  1098

Table 2
Properties of skin.

Parameters Value References

Thickness L (m) 0.006 [9]

Skin density r (kg/m3) 1190.0 [10]
blood density rb (kg/m3) 1060 [10]
Skin thermal conductivity k (W/m-K) 0.235 [11]
Skin specic heat c (J/kg-K) 3600.0 [12]
Metabolic heat generation qmet (W/m3) 368.1 [19]
Ratio of the thermal conductivity to the heat capacity, j k=rcL2 (sec) 0.001524 Based on the above parameters

Zt
Ut A exp  Ea =RTdt
0 temperature of skin is assumed to be T(x, 0) 37  C. The surface
where the frequency factorA 3.1  1098. Ea is the activation energy
and R is the universal gas constant. The ratio is described as Ea/
R 75000. It is now widely accepted that: U 0.53, the rst degree
burn; U 1.0, the second degree burn; U 104, the third degree
burn [14,22]. The activation energy Ea and the frequency factor A in
the model given by Henriques and Moritz [27] are independent to
the temperature. If the temperature is xed or steady state has been
achieved, thermal damage is linearly related to the exposure time.
However, it is observed that at elevated temperatures around
40e41  C cells are inactivated. This inactivation is temporary and
after some time the cells become resistance to heat. This phe-
nomenon is called as thermotolerance [28]. If the steady state
temperature is higher than the threshold value of 42  C, thermal
damage becomes linearly related to the exposure time. For hyper-
thermia treatment, temperature of tumor needs to be raised to at
least 42  C [28,29].
Moreover, several models for prediction of skin burn injury were
developed by empirically tting to experimental data. The value of
the activation energy Ea and the frequency factor A were adjusted to
obtain an optimal match over the range of states for which data
were available. the activation energy and the frequency factor of
several different damage models depend on the temperature
[30e35]. These models are listed in Table 1.
6. Numerical results and discussion
In general, the initial temperature of skin is described as T(x,
0) 37  C. The properties of the skin are presented in Table 2. All
properties of skin are assumed to independent to the temperature.
One investigates the effects of the boundary conditions, conduction
models, material properties on the temperatures variation and
thermal damage as follows:
Considering the constant temperature of body, Xu et al. [14]
assumed that the temperature of the inner side of skin is con-
stant, i.e.T(L, t) 37  C. But Ahmadikiaet al. [19] assumed the
isolated boundary condition at the inner side of skin, i.e. vT(L, t)/
vx 0. The effects of the two different boundary conditions at x L
(46) and thermal wave propagation are investigated rstly. The initial
heating of constant temperature is considered here. The corre-
sponding boundary condition at the outer side of skin is [21]

T0; t 1001  ut  15 C; t >0 (47)
where u is the step function. It is like a square temperature wave is
inputted into skin at x 0. Fig. 1a shows the effect of boundary
conditions and heat conduction models on the temperature varia-
tion. It is found that, for this case with the blood perfusion rate
6b 0.1, the temperature response of the skin at T(L,t) 37  C is
close to that at vT(L,t)/vx 0. Moreover, the phenomenon of ther-
mal wave propagation is signicant in the C-V model. When the
incident square temperature wave is given at x 0, the tempera-
ture variation at x 1 mm is similar to the square temperature
wave in the C-V model. The wave speed is about 1/13(mm/s).
However, the wave propagation is not obvious in the Pennes and
DPL models. The response rate in the DPL model is signicantly
slower than that in the Pennes model. In addition, their tempera-
ture variations are greatly different to that in the C-V model. In the
Pennes model, when the boundary temperature is abruptly
increased to 100  C, the temperature at x 1 mm will be greatly
increased. It demonstrates that the temperature response in the
Pennes model to the boundary temperature change is faster than
those in the C-V and DPL models.
Fig. 1b demonstrates the effect of the phase lag for heat ux tq
on the temperature at x 1 mm in the C-V model. The larger the
phase lag for heat ux tq is, the more the square temperature
response is. Moreover, the larger the phase lag for heat ux tq is, the
slower the thermal wave speed c is. The relation between the phase
lag for heat ux tq and the thermal wave speed c is shown in Fig. 1c.
If the phase lag tq approaches zero, the thermal wave speed c is
innite. If the phase lag tq is increased, the thermal wave speed c is
signicantly decreased. It is well known [36] that if the phase lag tq
is excessively large, the thermal wave speed VT in the C-V model can
p
be approximated as VT k=tq rc. Fig. 1c demonstrates that in the
C-V model, if the phase lag tq is large, the presented results are
152 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158

Fig. 1. aTemperature variation at x 1 mm in different models [6b 0:1, T(0,0 < t < 15s) 100  C,T(0,15s < t < 60s) 0  C; BC1: T(L,t) 37  C; BC2: vT(L,t)/vx 0]. b Effect of the
thermal lag for heat ux tq on the temperature variation at x 1 mm in the C-V model [6b 0:1, T(0,0 < t < 15s) 100  C,T(0,15s < t < 60s) 0  C; BC1: T(L,t) 37  C]. c Effect of
the thermal lagstq and tT on the wave speed [6b 0:1, T(0,0 < t < 15s) 100  C,T(0,15s < t < 60s) 0  C; BC1: T(L,t) 37  C]. d Temperature distributions of skin in different models
[6b 0:1,T(0,0 < t < 15s) 100  C, T(0,15s < t < 60s) 0  C; BC1: T(L,t) 37  C; BC2: vT(L,t)/vx 0]. e Effect of the blood perfusion rate 6b on temperature distribution of skin at
different time [tq tT 1,T(0,0 < t < 150s) 100  C, T(0,150s < t < 600s) 0  C; BC1: T(L,t) 37  C; BC2: vT(L,t)/vx 0]. f Effect of the blood perfusion rate 6b on temperature
variation at x 3 mm [tq tT 1,T(0,0 < t < 150s) 100  C, T(0,150s < t < 600s) 0  C; BC1: T(L,t) 37  C; BC2: vT(L,t)/vx 0].
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158 153

Fig. 2. a Diagram of pulse heat ux heating and uid cooling at the skin surface, x 0. b Temperature variation at x 0 due to the pulse laser heating and uid cooling in different
models [p 5w/cm2, 6b 0:1,Nu0 2.55,T 25  C,tperiod 2s,tpulsed width 1s,npulsed 5,L 6 mm, T(L,t) 37  C]. c Temperature response at x 1 mm to the pulse laser heating
and uid cooling in different models [p 5w/cm2, 6b 0:1,Nu0 2.55, T 25  C, tperiod 2s,tpulsed width 1s, npulsed 5,L 6 mm, T(L,t) 37  C]. d Temperature distribution of
skin due to the pulse laser heating and uid cooling in different models [p 5w/cm2, 6b 0:1,NuL 2.55, T 25  C, tperiod 2s,tpulsed width 1s, npulsed 5,L 6 mm, T(L,t) 37  C].
e Thermal damage index U at x 0 due to the pulse laser heating and uid cooling in different models [p 5w/cm2, 6b 0:1,Nu0 2.55, T 25  C, tperiod 2s,tpulsed width 1s,
npulsed 5, L 6 mm, T(L,t) 37  C]. f Thermal damage distribution due to the pulse laser heating and uid cooling in different models at 40 s after laser heating [p 5w/cm2,
6b 0:1,Nu0 2.55, T 25  C, tperiod 2s,tpulsed width 1s, npulsed 5, L 6 mm, T(L,t) 37  C].
154 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158
Fig. 3. a Effect of temperature of cooling uidT on the temperature variation at x 0
under the pulse heat ux heating [p 5w/cm2, 6b 0:1,Nu0 2.55,
tperiod 2s,tpulsed width 1s, npulsed 5,L 6 mm, T(L,t) 37  C]. b Effect of the tem-
perature of cooling uidT on the thermal damage index U at x 0 under the pulse
heat ux heating [p 5w/cm2, 6b 0:1,Nu0 2.55,tperiod 2s,tpulsed width 1s,
npulsed 5,L 6 mm, T(L,t) 37  C].
consistent to those of the approximated formula. However, for
small phase lag tq, their wave speeds are distinguishable. Moreover,
if the phase lag tT is considered, the thermal wave speed VT is
notably higher than those of the approximated formula. In addition,
the larger the phase lag for temperature gradient tT is, the faster the
thermal wave speedVT is. Fig. 1d demonstrates that the effect of the
variation of boundary temperature at x 0 on the temperature the
skin in the Fourier model is remarkably faster than that in the DPL
model. The reason is that the thermal wave speed in the Pennes
model is innite. Fig. 1e demonstrates that, for the blood perfusion
rate 6b 0.1, the temperature response of the system with the
Fig. 4. a Effect of the Nusselt number Nu0 on the temperature variation at x 0 under
the pulse heat ux heating [p 5w/cm2, 6b 0:1, T 25  C,
tperiod 2s,tpulsed width 1s, npulsed 5,L 6 mm, T(L,t) 37  C]. b Effect of the Nusselt
number Nu0 on the thermal damage index U at x 0 under the pulse heat ux heating
[p 5w/cm2, 6b 0:1, T 25  C, tperiod 2s,tpulsed width 1s, npulsed 5,L 6 mm,
T(L,t) 37  C].
constant temperature condition T(L,t) 37  C is close to that with
the isolated boundary condition vT(L,t)/vx 0.For the blood
perfusion rate 6b 0.01, increasing the heat conduction time
signicantly makes the difference between the temperature re-
sponses of the systems with the constant temperature and isolated
conditions at x L. Further, if the blood perfusion rate 6b is small,
their temperature difference close to the boundary x L is enor-
mous. The reason is that the blood in the skin will take heat away
and decrease the effect of boundary heating on the inner skin
temperature. Therefore, if the blood perfusion rate 6b is large, the
temperature at T(L,t) 37  C is close to that at vT(L,t)/vx 0. It is
concluded that the smaller the blood perfusion rate 6b is, the more
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158 155

model, the pulse temperature response to the pulse laser heating is

signicant. It is because the thermal wave speed of the Pennes
model is innite. However, the temperatures in the DPL model
whentq tT 10 sec, is unresponsive to the pulse heating. Fig. 2c
demonstrates that the temperature variation at x 1 mm due to
the pulse laser heating is unresponsive, because it is far from the
heating boundary. Fig. 2d demonstrates that in the C-V model the
temperature distribution is not smooth because of the thermal
wave effect. The conventional thermal damage model given by
Henriques [29] is used to describe the damage of skin by laser
heating. Fig. 2e shows that the thermal damage index U of skin at
x 0 signicantly increases with the laser pulse heating. After the
heating, time the thermal damage index U will be constant. The
thermal damage index U in the Pennes model is signicantly higher
than those in the C-V and DPL models. Fig. 2f shows that the shorter
the distance from the skin surface is, the higher the thermal
damage index U is.
Fig. 3a demonstrates the effect of temperature T of cooling
uid on the temperature of skin subjected to pulse laser heating. It
is found that although the cooling temperature T is decreased to
zero, in the Pennes model the pulse temperature response to the
pulse laser heating is signicant. The highest temperature by
T 0  C is lower by about 4  C than those by T 25  C.Fig. 3b
demonstrates that the thermal damage index U(T 0  C) will be
decreased to U(T 25  C)/10. It is concluded that decreasing the
cooling temperature T will signicantly decrease the skin tem-
perature and the thermal damage.
Fig. 4a demonstrates the effect of the Nusselt number Nu0 of
cooling convection on the temperature of skin subjected to pulse
laser heating. It is found that, for different Nusselt number Nu0, the
pulse temperature response to the pulse laser heating is always
signicant in the Pennes model. The highest temperature with
Nu0 2.55 is lower by about 14  C than those with Nu0 0.255 in
the Pennes model and 6.5  C in the DPL model. Further, the highest
temperature with Nu0 100 is lower by about 80  C than those with
Nu0 0.255 in the Pennes model and 42  C in the DPL model. Fig. 4b
demonstrates that in the Pennes model, the thermal damage index
U(Nu0 2.55) is equal toU(Nu0 0.255)/1000.In the DPL model,

Fig. 5. a Effect of the way of pulse laser heating {tpulsed width,tperiod,npulse,H} on the
temperature variation at x 0 [6b 0:1,Nu0 2.55,T 25  C,L 6 mm,
T(L,t) 37  C]. b Effect of the way of pulse laser heating {tpulsed width,tperiod,npulse,H} on
the thermal damage index U at x 0 [6b 0:1,Nu0 2.55,T 25  C,L 6 mm,
T(L,t) 37  C].

the difference between the temperature responses of the systems

with different conditions at the inner side of skin. It is also veried
in Fig. 1f. In reality, the perfusion rate is
0:0002 1=s < 6b < 0:00057 1=s [37e40]. Because the perfusion
rate of tissue is very small, the assumption of the isolated boundary
condition vT(L,t)/vx 0 results in enormous error of temperature
response.
Secondly, the skin surface is assumed to be heated by pulse laser
and cooled by uid simultaneously. Fig. 2a shows the diagram of
pulse heat ux heating and uid cooling at the skin surface x 0.
tpulsed width is the time per laser heating pulse. tperiod is the period
per laser pulse. npulse is the number of laser pulse and p is the heat Fig. 6. Comparison of thermal damage index U of several damage models at x 0 and
ux of laser heating pulse. Fig. 2b demonstrates that in the Pennes t 40s, due to pulsed laser heating and uid cooling [6b
0:1,Nu0 2.55,T 25  C,tperiod 2s,tpulsed width 1s,npulsed 5,L 6mm,T(L,t) 37  C].
156 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158
Fig. 7. Temperature variation of skin subjected to the pulsed heating and simultaneous
cooling for the thermal therapy; (a) the Pennes model, (b) the C-V model, (c) the DPL
model [p 1w=cm2 ; Nu0 2:55; tpulsed width 12tperiod ; tperiod 2; T 00 c].
Fig. 8. Effect of the heat transfer per unit area of each pulsed laser heating p on the
uctuation of the skin temperature at x 0
[Nu0 2:55; tpulsed width 12tperiod ; tperiod 2 sec; T 00 c].
Fig. 9. Effect of the period of pulsed heating tperiod on the uctuation of the skin
temperature at x 0 [p 1w=cm2 ; Nu0 2:55; tpulsed width 12tperiod ; T 0].
U(Nu0 2.55) is equal to U(Nu0 0.255)/100. It is concluded that
increasing the heat convective efciency will signicantly decrease
the skin temperature and the thermal damage.
Fig. 5 demonstrates the skin response to four cases of pulse laser
heating; case 1:
ftpulsed width ; tperiod ; npulse ; Hg f0:5s; 2s; 5; 7:5 W  s=cm2 g; case
2: f1s; 2s; 5; 15 W  s=cm2 g;case 3: f1:5s; 2s; 5; 22:5 W  s=cm2 g;
case 4: f0:5s; 1s; 10; 15 W  s=cm2 g, where the total heat transfer H
 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158
Fig. 10. Effect of the temperature of cooling liquid on the highest and lowest tem-
perature of skin at x 0 under the pulsed
[p 1w=cm2 ; Nu0 2:55 tpulsed width 1s; tperiod 2s].
from laser is H npulsetpulsed widthp. Because the total heat transfer H
of case 3 is highest, the skin temperature is highest. The pulse
heating method of case 2 is different to that of case 4, but their total
heat transfer H from laser is identical. In the Pennes model, their
temperatures responses of cases 2 and 4 are different. However,
there is no difference in the DPL model. Fig. 5b demonstrates that
although the total heat transfer H of cases 2 and 4 are identical, the
thermal damage index U(case 4) is about U(case 2)/10 due to
different heating methods in the Pennes model. But their thermal
damage index are consistent in the DPL model when tq tT 10s. It
is concluded that the effect of pulse laser heating method on the
skin temperature and the thermal damage is signicant in the
Pennes model but minor in the DPL model when tq tT 10s.
Comparison of thermal damage index U of several damage
models at x 0 and t 40s, due to pulsed laser heating and uid
cooling is made and shown in Fig. 6. It is found that in spite of the
Pennes, C-V and DPL models, under the same condition, the dam-
age index U of the Wu model [35] is the highest and that of the
Fugitt model [32] is the lowest. It is concluded that the damage
index U of the Henriques model is overestimated because the
activation energy and the frequency factor are independent to the
temperature.
Thermal therapy covers all therapeutic treatments based on the
transfer of thermal energy into a part of or whole of the body [25].
According to temperature level and time duration, Habashet al. [41]
categorized into three different modalities: (1) Diathermia (heating
upto 41  C, for 6e72 h), (2) Hyperthermia (41e45  C, for
15e60 min), and (3) Thermal ablation (above 45  C, for 4e6 min).
Fig. 7 demonstrates the temperature uctuation of skin sub-
jected to the pulsed heating and simultaneous cooling for the
thermal therapy. It is found that at the beginning of heating the
average temperature will increase gradually and reach steady
temperature uctuation nally. The time from the transient state to
the steady state is the shortest in the Pennes model and that is the
longest in the DPL model. The temperature uctuation of skin in the
Pennes model is greatest and smallest in the DPL model.
157
Fig. 8 demonstrates the effect of the heat transfer per unit area
of each pulsed laser heating, p on the uctuation of skin tempera-
ture. It is found that the uctuation of temperature is proportional
to the pulsed heat transfer p.
Fig. 9 demonstrates the effect of the period of pulsed heating
tperiod on the uctuation of skin temperature. It is found that the
larger the period tperiod is, the greater the uctuation of temperature
is especially in the Pennes model. If the period tperiod is large
enough, the variation of temperature in the three models is xed
and same.
Fig. 10 demonstrates the effect of the temperature of cooling
liquid on the highest and lowest temperature of skin. It is found
that the higher the temperature of the cooling liquid is, the higher
the average temperature of skin is. The effect of the cooling tem-
perature on the uctuation of skin temperature is negligible.
Moreover, one can match the condition of hyperthermia therapy
(41e450c for 15e60min) by considering the cooling temperature
T 5  Ce22  C under the pulsed heating in the DPL model when
tq tT 10s.
7. Conclusion
In this paper, the formulation of the heat transfer for skin tissue
with general boundary conditions is made. The thermal damage
and therapy under the pulse heating in the Pennes, C-V and Dual-
heating
phase-lag models are investigated. The comparison of several
thermal damage models is made. The main phenomena are
revealed as follows:
a. In the C-V model, if the phase lag tq is large, the presented wave
speedpare consistent
to those of the approximated formula
VT k=tq rc. However, for small phase lag tq, the wave speed
via the approximated formula is underestimated. Moreover, if
the phase lag tT is considered, the thermal wave speed VT is
greatly higher than those of the approximated formula.
b. The larger the phase lag for temperature gradient tT is, the faster
the thermal wave speedVT is.
c. In the Pennes model, the pulse temperature responses to the
pulse laser heating is signicant. It is because the thermal wave
speed of the Pennes model is innite. However, the temperature
in the DPL model is unresponsive to the pulse heating.
d. If the Nusselt number Nu0 is small, the cooling effect is minor.
The smaller the Nusselt number Nu0 is, the more signicant the
pulse temperature response is.
e. The damage index U of the Henriques model is overestimated
because the activation energy and the frequency factor are in-
dependent to the temperature.
f. The larger the pulsed heat transfer p is, the greater the uctu-
ation of temperature is.
g. The larger the period tperiod of pulsed heating is, the greater the
uctuation of temperature is.
References
[1] Afrin N, Zhang Y, Chen JK. Thermal lagging in living biological tissue based on
nonequilibrium heat transfer between tissue, arterial and venous bloods. Int J
Heat Mass Transf 2011;54:2419e26.
[2] Pennes H. Analysis of tissue and arterial blood temperatures in the resting
human forearm. J Appl Physiol 1948;1:93e122.
[3] Liu J, Chen X, Xu L. New thermal wave aspects on burn evaluation of skin
subjected to instantaneous heating. IEEE Trans Biomed Eng 1999;46(4):
420e8.
[4] Antaki P. New interpretation of non-Fourier heat conduction in processed
meat. J Heat Transf 2005;127(2):189e93.
[5] Zhou J, Chen JK, Zhang Y. Dual-phase lag effects on thermal damage to bio-
logical tissues caused by laser irradiation. Comput Biol Med 2009;39:286e93.
[6] Catteneo C. A form of heat conduction equation which eliminates the paradox
of instantaneous propagation. Compte Rendus 1958;247:431e3.
158 S.-M. Lin, C.-Y. Li / International Journal of Thermal Sciences 110 (2016) 146e158
[7] Vernotte P. Les paradoxes de la theorie continue de lequation de la chaleur.
Compte Rendus 1958;246:3154e5.
[8] Vedavarz A, Kumar S, Moallemi MK. Signicance of non-Fourier heat waves in
conduction. ASME J Heat Transf 1994;116(1):221e4.
[9] Brazhnikov AM, Karpychev VA, Luikova AV. One engineering method of
calculating heat conduction processes. Inzhenerno Fiz Zh 1975;28(4):677e80.
[10] Kaminski W. Hyperbolic heat conduction equation for materials with a
nonhomogeneous inner structure. ASME J Heat Transf 1990;112(3):555e60.
[11] Mitra K, Kumar S, Vedavarz A, Moallemi MK. Experimental evidence of hy-
perbolic heat conduction in processed meat. ASME J Heat Transf 1995;117(3):
568e73.
[12] Stratigos AJ, Dover JS. Overview of lasers and their properties. Derm Ther
2000;13:2e16.
[13] Xu F, Seffen K a, Lu TJ. Non-Fourier analysis of skin biothermomechanics. Int J
Heat Mass Transf 2008;51:2237e59.
[14] Tzou D. A unied eld approach for heat conduction from macro-to micro-
scales. J Heat Transf 1995;117(1):8e16.
[15] Xu F, Lu TJ, Seffen KA. Biothermomechanics of skin tissues. J Mech Phys Solids
2008;56:1852e84.
[16] Ng EYK, Chua LT. Prediction of skin burn injury. Part 1: numerical modelling,
Source. Proc Inst Mech Eng Part H J Eng Med 2002;216(H3):157e70.
[17] Deng Z-S, Liu J. Analytical study on bioheat transfer problems with spatial or
transient heating on skin surface or inside biological bodies. J Biomech Eng
2002;124:638e49.
[18] Ozen S, Helhel S, erezci O. Heat analysis of biological tissue exposed to mi-
crowave by using thermal wave model of bio-heat transfer (TWMBT). Burns
2008;34:45e9.
[19] Ahmadikia H, Fazlali R, Moradi A. Analytical solution of the parabolic and
hyperbolic heat transfer equations with constant and transient heat ux
conditions on skin tissue. Int Commu Heat Mass Transf 2012;39:121e30.
[20] Xu F, Lu TJ, Seffen K a, Ng EYK. Mathematical modeling of skin bioheat
transfer. Appl Mech Rev 2009;62:050801.
[21] Liu KC, Wang YN, Chen YS. Investigation on the bio-heat transfer with the
dual-phase-lag effect. Int J Therm Sci 2012;58:29e35.
[22] Lin SM. Analytical solutions of bio-heat conduction on skin in fourier and non-
fourier models. J Mech Med Biol 2013;13(4):1350063.
[23] Kumar P, Kumar D, Rai KN. A numerical study on dual-phase-lag model of bio-
heat transfer during hyperthermia treatment. J Therm Biol 2015;49e50:
98e105.
[24] Hooshmand P, Moradi A, Khezry B. Bioheat transfer analysis of biological
tissues induced by laser irradiation. Int J Therm Sci 2015;90:214e23.
[25] Kumar D, Singh S, Rai KN. Analysis of classical fourier, SPL and DPL heat
transfer model in biological tissues in presence of metabolic and external heat
source. Heat Mass Transf 2015. http://dx.doi.org/10.1007/s00231-015-1617-0.
[26] Dombrovsky, Timchenko VM. Laser induced hyperthermia of supercial tu-
mors: computational models for radiative transfer, combined heat transfer,
and degradation of biological tissues. Therm Process Eng 2015;7(1):24e36 [in
Russian)].
[27] Henriques FC, Moritz AR. Studies of thermal injury, 1. The conduction of heat
to and through skin and the temperatures attained therein. A theoretical and
an experimental investigation. Am J Pathol 1947;23:531e49.
[28] Habash RWY, Bansal R, Krewski D, Alhad HT. Thermal therapy. Part 2: hy-
perthermia techniques. Crit Rev Biomed Eng 2006;34:491e542.
[29] Jamil M, Ng EYK. To optimize the efcacy of bioheat transfer in capacitive
hyperthermia: a physical perspective. J Therm Biol 2013;38(5):272e9.
[30] Ng EYK, Chua LT. Prediction of skin burn injury. Part 2: parametric and
sensitivity analysis. Proc Inst Mech Eng Part H J Eng Med 2002;216:171e83.
[31] Henriques FC, Studies of thermal injury. V. The predictability and the signif-
icance of thermally induced rateprocesses leading to irreversible epidermal
injury. Arch Pathol 1947b;43:489e502.
[32] Fugitt CE. A rate process of thermal injury. Armed ForcesSpecial Weapons
Project; 1955. AFSWP-606.
[33] Stoll AM, Greene LC. Relationship between pain and tissue damage due to
thermal radiation. J Appl Phys 1959;14(3):373e82.
[34] Takata AN. Development of criterion for skin burns. Aerosp Med 1974;45:
634e7.
[35] Wu YCA. Modied criterion for predicting Thermal Injury. Washington: Na-
tional Bureau of Standards; 1982.
[36] Joseph DD, Preziosi L. Heat waves. Rev Mod Phys 1989;61(1):41e73.
[37] Jamil M, Ng EYK. Ranking of parameters in bioheat transfer using Taguchi
analysis. Int J Therm Sci 2013;63:15e21.
[38] Jamil M, Ng EYK. Evaluation of meshless radial basis collocation method
(RBCM) for heterogeneous conduction and simulation of temperature inside
the biological tissues. Int J Therm Sci 2013;68:42e52.
[39] Jamil M, Ng EYK. Statistical modeling of electrode based thermal therapy with
Taguchi based multiple regression. Int J Therm Sci 2013;71:283e91.
[40] Jamil M, Ng EYK. Parametric sensitivity analysis of radiofrequency ablation
with efcient experimental design. Int J Therm Sci 2014;80:41e7.
[41] Habash RWY, Bansal R, Krewaski D, Alhad HT. Thermal therapy part-I: an
introduction to thermal therapy. Crit Rev Biomed Eng 2006;34(6):459e89.