energies

Article
Electrothermal Model of SiC Power BJT
Joanna Patrzyk *, Damian Bisewski and Janusz Zar˛ebski
Department of Marine Electronics, Gdynia Maritime University, 81-225 Gdynia, Poland;
d.bisewski@we.umg.edu.pl (D.B.); j.zarebski@we.umg.edu.pl (J.Z.)
* Correspondence: j.patrzyk@we.umg.edu.pl; Tel.: +48-58-5586-575
Article



Received: 28 March 2020; Accepted: 17 May 2020; Published: 21 May 2020

Electrothermal Model of SiC Power BJT
Abstract: This paper refers to the issue of modelling characteristics of SiC power bipolar junction
Joanna Patrzyk *, Damian Bisewski and Janusz Zarębski
transistor (BJT), including the self-heating phenomenon. The electrothermal model of the tested
Department of Marine Electronics, Gdynia Maritime University, 81-225 Gdynia, Poland;
device is demonstrated and experimentally verified. The electrical model is based on the isothermal
d.bisewski@we.umg.edu.pl (D.B.); j.zarebski@we.umg.edu.pl (J.Z.)
Gummel–Poon model, but several
* Correspondence: modifications were
j.patrzyk@we.umg.edu.pl; made including the improved current gain
Tel.: +48-58-5586-575
factor (β) model and the modified
Received: 28 March 2020; model of the
Accepted: quasi-saturation
17 May region.
2020; Published: 21 May 2020 The accuracy of the presented
model was assessed by comparison of measurement and simulation results of selected characteristics
Abstract: This paper refers to the issue of modelling characteristics of SiC power bipolar junction
of the BT1206-ACtransistor
SiC BJT manufactured by TranSiC. In this paper, a single device characterization
(BJT), including the self-heating phenomenon. The electrothermal model of the tested device
has only been performed.
is demonstrateddemonstrated
The and experimentally results of research
verified. showmodel
The electrical the evident
is based temperature impact
on the isothermal
on the transistor d.c. characteristics. A good compliance between the measured and calculated
Gummel–Poon model, but several modifications were made including the improved current gain
characteristics of factor (β) model and the modified model of the quasi-saturation region. The accuracy of the presented
the considered transistor is observed even in quasi-saturation mode.
model was assessed by comparison of measurement and simulation results of selected characteristics
of the BT1206-AC SiC BJT manufactured by TranSiC. In this paper, a single device characterization
Keywords: BJT; modelling;
has only been self-heating; silicon carbide;
performed. The demonstrated resultsSPICE
of research show the evident temperature impact
on the transistor d.c. characteristics. A good compliance between the measured and calculated
characteristics of the considered transistor is observed even in quasi-saturation mode.

Keywords: BJT; modelling; self-heating; silicon carbide; SPICE

1. Introduction
While the power MOSFET made of silicon carbide (SiC) is currently the most popular switching
1. Introduction
power wide bandgap-based semiconductor device, the SiC bipolar junction transistor (BJT) is still
While the power MOSFET made of silicon carbide (SiC) is currently the most popular switching
an inviting counterpart dedicated to high-temperature/high-power systems, due to its particular
power wide bandgap-based semiconductor device, the SiC bipolar junction transistor (BJT) is still an
properties, such as low specific
inviting counterpart on-resistance,
dedicated to a less complicated manufacture
high-temperature/high-power systems, method
due to itsand its being
particular
free of oxide reliability referred to the high value of the electrical field and temperature. Sincebeing
properties, such as low specific on-resistance, a less complicated manufacture method and its the last
decade, the market freehas
of oxide
beenreliability
offering referred
high to the highsilicon
voltage value ofcarbide
the electrical
powerfield and
BJTstemperature.
[1–5]. Since the last
decade, the market has been offering high voltage silicon carbide power BJTs [1–5].
The research subject in this paper is the high voltage BJT made of silicon carbide, BT1206-AC
The research subject in this paper is the high voltage BJT made of silicon carbide, BT1206-AC
manufactured bymanufactured
the TranSiCbycompany
the TranSiC and placed
company andinplaced
a typical TO-247
in a typical TO-247case. The
case. Thevalues
values ofof the
the certain
certain
parameters of thisparameters
transistor are transistor
of this demonstrated in Tablein1 Table
are demonstrated [5]. 1 [5].
Table 1. Parameter values of the BT1206-AC transistor.
Table 1. Parameter values of the BT1206-AC transistor.
VCEO ICmax Tjmax
Manufacturer Type ◦ C] β [A/A] Package
Manufacturer Type VCEO [V] [V] [A] [A]Tjmax [[°C]
ICmax β [A/A] Package
TO-247
TO-247
BT1206- 35 at
35 at
TranSiC TranSiC
BT1206-AC 1200 1200
6 6 175175
AC (V CE =
(VCE = 2.5
2.5 V, =44A)
C =
V, IC A)

Due to the improved electrical and thermal properties of silicon carbide devices compared to
Due to the improved electrical
silicon devices, companiesandstate
thermal properties
that designers willofwant
silicon carbide
to replace devices
silicon compared
components for to
applications such as hybrid electric vehicles, inverters of solar cells and high
silicon devices, companies state that designers will want to replace silicon components for applicationspower electric motor
controllers with their silicon carbide counterparts. For all these applications, silicon carbide can solve
such as hybrid electric vehicles, inverters of solar cells and high power electric motor controllers with
the problems with a high-temperature environment and switching losses.
their silicon carbide counterparts. For all these applications, silicon carbide can solve the problems
with a high-temperature environment and switching losses.
Energies 2020, 13, x; doi: FOR PEER REVIEW www.mdpi.com/journal/energies

Energies 2020, 13, 2617; doi:10.3390/en13102617 www.mdpi.com/journal/energies

 Energies 2020, 13, 2617 2 of 9
Energies 2020, 13, x FOR PEER REVIEW 2 of 9

The ambient temperature has a strong impact on certain characteristics and parameters of the
The ambient temperature has a strong impact on certain characteristics and parameters of the
silicon carbide power bipolar transistors [4,6–8]. Moreover, in real operating conditions, due to the
silicon carbide power bipolar transistors [4,6–8]. Moreover, in real operating conditions, due to the
self-heating phenomenon, the junction temperature exceeds the ambient temperature because of the
self-heating phenomenon, the junction temperature exceeds the ambient temperature because of the
device dissipated thermal power converted into the heat in non-ideal heat dissipation conditions. The
device dissipated thermal power converted into the heat in non-ideal heat dissipation conditions.
self-heating phenomenon, resulting in qualitative and quantitative changes of the semiconductor
The self-heating phenomenon, resulting in qualitative and quantitative changes of the semiconductor
device characteristics shape, is observed as well in SiC BJTs [6].
device characteristics shape, is observed as well in SiC BJTs [6].
In order to properly analyze the operation of SiC BJT-based electronic systems, a precise and
In order to properly analyze the operation of SiC BJT-based electronic systems, a precise and
relatively simple SiC BJT model for circuit simulation and trend analysis is needed. One of the earliest
relatively simple SiC BJT model for circuit simulation and trend analysis is needed. One of the earliest
semi-physical models for silicon BJT is the Gummel and Poon model (G–P model) proposed in 1970
semi-physical models for silicon BJT is the Gummel and Poon model (G–P model) proposed in 1970 [9].
[9]. The G–P model has become a fundamental model for characteristics calculation of BJT in SPICE
The G–P model has become a fundamental model for characteristics calculation of BJT in SPICE
(Simulation Program with Integrated Circuit Emphasis) which is a popular computer tool that allows
(Simulation Program with Integrated Circuit Emphasis) which is a popular computer tool that allows
the analysis of characteristics and parameters of electronic components and systems [7,10–14].
the analysis of characteristics and parameters of electronic components and systems [7,10–14]. Certain
Certain modifications have been made to evolve SiC BJT models in recent years. In 2004, the G–P
modifications have been made to evolve SiC BJT models in recent years. In 2004, the G–P model
model was applied to calculate the characteristics of SiC BJT for the first time [15]. Johannesson and
was applied to calculate the characteristics of SiC BJT for the first time [15]. Johannesson and Nawaz
Nawaz established a G–P-based analytical SPICE model for SiC BJT power modules that is useful for
established a G–P-based analytical SPICE model for SiC BJT power modules that is useful for the
the designing of power electronics systems containing considered devices [16]. However, these
designing of power electronics systems containing considered devices [16]. However, these models do
models do not take into account the self-heating phenomenon. There are some electrothermal models
not take into account the self-heating phenomenon. There are some electrothermal models (ETMs)
(ETMs) dedicated to SiC BJT transistors in the literature [6,17,18], but works describing these models
dedicated to SiC BJT transistors in the literature [6,17,18], but works describing these models do not
do not contain full information enabling their implementation in the SPICE program (e.g., no
contain full information enabling their implementation in the SPICE program (e.g., no description of
description of the controlled sources used in the model) or use numerical models that are complex,
the controlled sources used in the model) or use numerical models that are complex, demand specific
demand specific data concerning material features or device geometry and have time-consuming
data concerning material features or device geometry and have time-consuming simulations.
simulations.
This paper refers to the issue of modelling of static characteristics of SiC power BJTs including the
This paper refers to the issue of modelling of static characteristics of SiC power BJTs including
self-heating phenomenon. The electrothermal model of SiC BJT has been proposed. The model was
the self-heating phenomenon. The electrothermal model of SiC BJT has been proposed. The model
experimentally examined by comparison of the measured and simulated isothermal and non-isothermal
was experimentally examined by comparison of the measured and simulated isothermal and non-
characteristics of the selected transistor. The effect of self-heating on transistor characteristics was
isothermal characteristics of the selected transistor. The effect of self-heating on transistor
assessed and discussed. In this paper, the experimental verification of the model has been performed
characteristics was assessed and discussed. In this paper, the experimental verification of the model
using a single device characterization, however, future work will include verification results based on
has been performed using a single device characterization, however, future work will include
a larger population of the considered class of devices.
verification results based on a larger population of the considered class of devices.
2. The Model Form
2. The Model Form
A block diagram of the electrothermal model (ETM) of SiC BJT proposed by the authors is
A block diagram
demonstrated in Figureof1.the electrothermal model (ETM) of SiC BJT proposed by the authors is
demonstrated in Figure 1.
C V, I

B ELECTRICAL HEAT POWER

MODEL MODEL
ETM
E PTH
Tj
THERMAL
MODEL

Ta
Figure 1. A block diagram of the electrothermal model of SiC bipolar junction transistor (BJT).
Figure 1. A block diagram of the electrothermal model of SiC bipolar junction transistor (BJT).
As presented, the ETM contains: the electrical model, the heat power model and the thermal
As The
model. presented, the
function of ETM contains:
the electrical the electrical
model model,the
is to determine the heatcharacteristics
static power modelofand the thermal
the transistor for
model. The function of the electrical model is to determine the static characteristics of the
a known temperature (Tj ) of its interior. The heat power model is designed to determine the value transistor
for a known temperature (Tj) of its interior. The heat power model is designed to determine the value
of electrical power generated in the transistor based on the currents and clamping voltages values on
Energies 2020, 13, 2617 3 of 9

of Energies
electrical power generated in the transistor based on the currents and clamping voltages 3values
2020, 13, x FOR PEER REVIEW of 9
on the transistor leads. The thermal model calculates the actual value of transistor junction (inner)
temperature
the transistor based on the
leads. Theambient
thermal temperature (Ta ) and
model calculates the thermal
the actual value parameters
of transistorassociated with the
junction (inner)
transistor
temperaturecooling
based method.
on the ambient temperature (Ta) and the thermal parameters associated with the
The form
transistor of themethod.
cooling considered electrical model is based on the G–P model. However, in our model,
The form of thecollector
a voltage-dependent considered electrical
resistance is model
used foris based on thethe
modelling G–P model. However,
quasi-saturation in our
mode thatmodel,
starts to
beavisible
voltage-dependent collector resistance
in the higher temperature range inis SiC
usedBJTs’
for modelling thecharacteristics.
static output quasi-saturation mode
The that starts
quasi-saturation
to bemodel
region visibleisinbased
the higher
on thetemperature
Kull model range in SiC BJTs’
and consists of thestatic output characteristics.
temperature-dependent The quasi-
intrinsic carrier
saturation region model is based on the Kull model and consists of the temperature-dependent
concentration model dedicated to the devices made of silicon carbide. A network form of the proposed
intrinsicmodel
electrical carrierisconcentration
demonstratedmodel dedicated
in Figure to the devices made of silicon carbide. A network
2 [1,16,19].
form of the proposed electrical model is demonstrated in Figure 2 [1,16,19].
RC
C
C’
IBC1
IBC2 BR (IBE-IBC)
IN = Kqb
D1 D2 GN
B RB B’

D3 D4

IBE2 IBE1
BF
E’ RE
E
Figure2.2. Network
Figure form of
Network form of the
theelectrical
electricalmodel
modelofofSiC
SiCBJT.
BJT.

Diodes
DiodesDD1 –D
1–D4 4 model
model the
the generation
generation and and recombination
recombinationphenomena phenomenain inthethelayers
layers
of of
thethe
collector/emitter-base junctions, whereas
collector/emitter-base junctions, whereas D –D and
1 D21–D2 andD 3 –D D represent the ideal and non-ideal
4 3–D4 represent the ideal and non-ideal component
of component
these junctions, respectively.
of these junctions,Values of the collector
respectively. Values of and thethecollector
base current of the
and the transistor
base current are of the
of the
form [19]: are of the form [19]:
transistor
I +I
IB = BE1IBE1+ IBC1BC1
+ IBE2 + IBC2 (1)
IB = β + IBE2 + IBC2 (1)
β
IBE1 IBC1 IBC1
IC = − − − IBC2 (2)
KIqbBE1 K IBC1
qb IBC1
β
IC = − − − IBC2 (2)
Kqb Kqb β
where the currents IBE1 , IBE2 , IBC1 , IBC2 are defined as follows [19]:
where the currents IBE1, IBE2, IBC1, IBC2 are defined as follows [19]:
V
  
IBE1 = IS· exp VBE BE
− 1 (3)
IBE1 =IS· exp NF·Vt − 1 (3)
NF·Vt
V
   
IBE2 = ISE· exp VBE − 1 (4)
IBE2 =ISE· exp NE·Vt − 1
BE
(4)
 NE·Vt
VBC
 
IBC1 = IS· exp −1 (5)
NR·V
VBC t
IBC1 =IS· exp 
NR·V − 1  (5)
VBCt
IBC2 = ISC· exp −1 (6)
NC·Vt
VBC
where β is the current gain factor, VBE IisBC2 the=ISC·
intrinsic −1
exp (B’–E’) base-emitter (6)
voltage, VBC is the intrinsic
NC·Vt
(B’–C’) base-collector voltage, IS, ISE and ISC are the saturation currents, NF and NR are the current
where β is the current gain factor, VBE is the intrinsic (B’–E’) base-emitter voltage, VBC is the intrinsic
emission coefficients, NE and NC are the leakage emission coefficients and coefficient Kqb is of the
(B’–C’) base-collector voltage, IS, ISE and ISC are the saturation currents, NF and NR are the current
form [19]:
emission coefficients, NE and NC are the leakage" emission coefficients and coefficient
NK # Kqb is of the
V V I I
 
form [19]: Kqb = 0.5 · (1 − BC − BC )−1 · 1 + 1 + 4 · BE1 + BC1 (7)
VAF VAF IKF IKR
Energies 2020, 13, 2617 4 of 9

where VAF, and VAR are early voltages, IKF and IKR are the knee current parameters, and NK is the
high-current roll-off coefficient.
Efficiency of the controlled-current source GN from Figure 2 is of the form [19]:

IBE − IBC
IN = (8)
Kqb

The coefficient β existing in Equations (1) and (2) are defined as [20]:
    !
β0 · 1 + a· Tj −T0  X q 
2
β=    q ·1−
 γi · exp αi · IC + d1  (9)
1 + b· 1 + c· Tj −T0 · I2C +d1 i

where T0 is the nominal temperature, β0 , αi , γi, a, b, c, and d1 are the model parameters.
Resistors RE and RB represent series resistances of the emitter and base areas according to
Equation [19]:     2 
RE (T) = RE · 1 + TRE1· Tj − T0 + TRE2· Tj − T0 (10)
    2 
RB (T) = RB · 1 + TRB1· Tj − T0 + TRB2· Tj − T0 (11)

where TRE1, TRB1, TRE1 and TRE2 represent linear and quadratic temperature coefficients of the
emitter and the base resistances, respectively.
The voltage-dependent collector resistance (RC ) dedicated for the quasi-saturation region modeling
is described as [1,16]:
 s ! s !−1
 1 1 4·n 2 q·V 1 4·n 2 q·V
i B0C0 i B0C 
RC = RC0 · + · 1 + · exp + · 1 + · exp  (12)
2 4 Nepi 2 k·Tj 4 Nepi 2 k·Tj 


where Nepi is an epi-layer collector doping concentration, VB’C’ and VB’C are voltages in the selected
end of the collector resistance, q is the elementary charge, and ni is the intrinsic carrier concentration
defined as [1,16]:
Eg0
!
3
2
ni = A ·Tj · exp − (13)
2·k·Tj
where Eg0 is SiC bandgap at zero temperature, k is Boltzmann’s constant, A is the material constant
independent of temperature.
The zero-bias collector resistance RC0 existing in Equation (12) is of the form [1,16]:

RC0 = 2.143 · 10−21 · exp (0 .1302 · Tj (14)

A detailed specification of the mathematical description of the considered model is achievable in

the SPICE user manual [19] and other papers [1,16,20].
In the thermal model, the heat interactions between the device interior and the ambient temperature
were taken into account. As a rule, the compact thermal model of a semiconductor device is represented
by the form of the RC network. Based on this form from other works [21,22], the non-linear
compact thermal model of the SiC BJT was demonstrated. Thermal capacitances hardly depend on the
temperature in contrast to the thermal resistance that changes with the temperature. In the presented
model controlled voltage sources were used only instead of R elements. The network representation of
the considered model is shown in Figure 3.
 Energies 2020, 13, 2617 5 of 9
Energies 2020, 13, x FOR PEER REVIEW 5 of 9

Tj ER1 ER2 ERn Ta

GP C1 C2 Cn
VTa

Figure
Figure3.3.Network
Networkform
formofofthe
thenon-linear
non-linearthermal
thermalmodel
modelofofthe
theSiC
SiCBJT.
BJT.

Thecontrolled-current
The controlled-currentsource sourceGGPPexpresses
expressesthethepower
powerdissipated
dissipatedininthe
thedevice,
device,controlled
controlledvoltage
voltage
sources
sourcesEER1R1, E, R2E,R2
…,, . E. .Rn, express temperature
ERn express temperature changes between
changes everyevery
between part ofpart
the of
thermal dissipation
the thermal path.
dissipation
The
path.analytical
The analyticalform ofform theseofvoltage source efficiencies
these voltage takes into
source efficiencies consideration
takes changes in
into consideration the thermal
changes in the
resistance value expressing
thermal resistance the thermalthe
value expressing dissipation between thebetween
thermal dissipation specific the
construction parts. Capacitors
specific construction parts.
CCapacitors
1–Cn describe C1 –C thermal
n describe capacitances
thermal of every
capacitances construction
of every part. The
construction voltage
part. at
The nodes
voltage T
atj and
nodes TTa j
represents the transistor
and Ta represents junction temperature
the transistor and the ambient
junction temperature and thetemperature respectively.
ambient temperature respectively.
Efficiencies
Efficiencies of controlled-current
controlled-currentsource sourceGPGand
P and controlled-voltage
controlled-voltage sources
sources are ofare
theofform
the[21,22]:
form
[21,22]: ! !
in
ERn = Rth1 · exp in +Rth0 · in · di (15)
ERn = Rth1 · exp p0 + Rth0 · in · di (15)
p0
Gp = pth = IB ·VBE +IC · VCE (16)
Gp = p = IB ·VBE + IC · VCE (16)
where di is the quotient of Rn thermal resistance th and the transistor thermal resistance Rth , Rth0 , Rth1 ,
p
where
0 are the model parameters,
di is the quotient of Rn thermal and i is the current
n resistance and ofthe
thetransistor
source ERnthermal
. resistance Rth, Rth0, Rth1, p0
are the model parameters, and in is the current of the source ERn.
3. Results of Measurements and Simulations
3. Results of the
To test Measurements andpresented
accuracy of the Simulations
electrothermal model, static isothermal and non-isothermal
characteristics of the BT1206-AC
To test the accuracy SiC BJT
of the presented were measured
electrothermal and
model, compared
static to the
isothermal andresults of SPICE
non-isothermal
calculations. The
characteristics model
of the was implemented
BT1206-AC SiC BJT were in SPICE in the and
measured formcompared
of the subcircuit. A set ofofelectrical
to the results SPICE
calculations. The model was implemented in SPICE in the form of the subcircuit. A set ofparameters
parameter values of the ETM was used in the simulations. The values of the model electrical
were acquired
parameter valuesfrom theETM
of the measured characteristics
was used with theThe
in the simulations. usevalues
of theofproper parameters
the model estimation
parameters were
procedure described in [20,23]. The values of parameters occurring in
acquired from the measured characteristics with the use of the proper parameters estimation Equation (15) were acquired
from measurements
procedure described of in the transistor
[20,23]. transient
The values thermal impedance.
of parameters occurring in Equation (15) were acquired
Isothermal characteristics
from measurements of thetransient
of the transistor selected thermal
transistor were measured by means of Keithley Source
impedance.
Measure Unit type
Isothermal 2602 in order
characteristics of thetoselected
minimize the impact
transistor were of the self-heating
measured by meansphenomenon on the
of Keithley Source
measured characteristics of the tested device. During the measurements, a signal
Measure Unit type 2602 in order to minimize the impact of the self-heating phenomenon on the measured with a duty cycle
of less than 0.0001
characteristics and a device.
of the tested pulse duration
During the of measurements,
50 µs was used.a signal
The non-isothermal
with a duty cyclecharacteristics were
of less than 0.0001
performed by the “point by point” measurement method for the transistor operating
and a pulse duration of 50 μs was used. The non-isothermal characteristics were performed by the "point without any
cooling
by point"system under method
measurement the thermal steady-state.
for the During the
transistor operating measurements,
without any coolingthe case under
system temperature TC of
the thermal
the transistor was measured by a TM-2000 precision thermometer comprising a platinum
steady-state. During the measurements, the case temperature TC of the transistor was measured by a TM- Pt-100 sensor.
The precision
2000 temperature sensor was
thermometer glued on athe
comprising transistor
platinum casesensor.
Pt-100 by a heat
Theconductive
temperature AG Termoglue,
sensor was glued10 G.
on
The measured
the transistor case by and
a heatsimulated
conductiveisothermal static output
AG Termoglue, 10 G. characteristics of the BT1206-AC transistor
are presented
The measuredin Figure 4, where points
and simulated represent
isothermal staticthe results
output of measurements
characteristics of the and lines represent
BT1206-AC the
transistor
results of simulations using the proposed ETM.
are presented in Figure 4, where points represent the results of measurements and lines represent the
resultsAsofseen in Figure
simulations 4, good
using agreement
the proposed between the results of the isothermal measurements in
ETM.
comparison with the simulations was obtained even in the quasi-saturation region at both values of the
ambient temperature, which expresses a percentage relative error (absolute error quotient and exact
value) of less than 8%. The increase of the ambient temperature causes the collector current value drop
in the bipolar silicon carbide transistors, but it is commonly known from the literature that in the case
of the bipolar transistors made of silicon, this dependence is reversed [24,25].
 Energies 2020, 13, x FOR PEER REVIEW 6 of 9

7.5 iB= 100 mA

Ta= 25ºC
Energies 2020, 13, 2617 6 of 9
Energies 2020, 13, x FOR PEER6REVIEW 6 of 9

4.5

iC [A]
7.5 iB= 100 mA Ta= 150ºC
Ta= 25ºC
3
6
1.5
4.5

iC [A]
0 Ta= 150ºC
3 0 1 2 3 4 5 6
VCE [V]
1.5
Figure 4. Isothermal output characteristics of the BT1206-AC transistor.
0
As seen in Figure 4, 0 1
good agreement 2
between 3 results
the 4 of the isothermal
5 6 measurements in
comparison with the simulations was obtained even VCE [V]
in the quasi-saturation region at both values of
the ambient temperature, which expresses a percentage relative error (absolute error quotient and
exact value) ofFigure
less than
Figure 4.8%. The increase
4. Isothermal
Isothermal output of the ambient
outputcharacteristics
characteristics temperature
ofofthe
theBT1206-AC
BT1206-AC causes the collector current
transistor.
transistor.
value drop in the bipolar silicon carbide transistors, but it is commonly known from the literature
Measured
Asthe
that in seenand
case calculated
in of
Figure 4, good
the bipolar non-isothermal
agreement
transistors I-V
between
made characteristics
the results
of silicon, of of
this dependence thethe considered
isothermal SiC BJT transistor
measurements
is reversed [24,25]. in
comparison
are shown Measured with
in Figure 5.theThe
and simulations
calculated was obtained
investigations
non-isothermal even
were performed
I-V in theatquasi-saturation
characteristics fouroffixed values region
the considered at both
of SiC
the values
control
BJT of
current
transistor in
the
are ambient
the range shown intemperature,
of 10–150Figure
mA. 5.AsThe which
shown expresses
investigations
in Figures a4 percentage
were and 5, the relative
performed at fourerror
differences fixed (absolute
values of
between theerror
the quotient
control
results and
ofcurrent
isothermal
exact
in thevalue)
range of
and non-isothermal ofless
10–150thanmA.
measurements8%. AsTheshown
increase of theoutput
of theinstatic
Figures ambient temperature
4 andcharacteristics
5, causes
the differences the collector
iC (vbetween current
the results of
CE ) reach over 50% for a
value drop and
isothermal in the bipolar siliconmeasurements
non-isothermal carbide transistors,
of thebut it is
static commonly
output known from
characteristics iC(vCEthe literature
) reach over
control current equal to iB = 100 mA and a voltage vCE of 4 V at room temperature.
that
50% in
forthe case of current
a control the bipolar
equal transistors
to iB = 100made
mA andof silicon,
a voltagethisvdependence
CE of 4 V at roomis reversed [24,25].
temperature.
Measured and calculated non-isothermal I-V characteristics of the considered SiC BJT transistor
are shown in Figure 5. The 5 investigations were performed at four fixed values of the control current
Ta= 25ºC
in the range of 10–150 mA. As shown in Figures 4 and 5, the differences between the results of
isothermal and non-isothermal 4 measurements of the static output characteristics iC(vCE) reach over
50% for a control current equal to iB = 100 mA and a voltageiB=vCE 150ofmA
4 V at room temperature.
3
iC [A]

5
Ta= 25ºC iB= 100 mA
2
4 iB= 50 mA

1 iB= 150 mA
3
iC [A]

iB= 10 mA
0 iB= 100 mA
2 0 2 4 6 8 10
i = 50 mA
VCE [V]B
1
Figure 5. Non-isothermal output characteristics of the BT1206-AC transistor.
iB= 10 of
Figure 5. Non-isothermal output characteristics mAthe BT1206-AC transistor.
0
As seen, the satisfying agreement
As seen, the satisfying 0 agreement2 between
between 4 the results
the results6 ofofsimulations
simulations
8 and
and
10 measurements was
measurements was
expressed by a relative error value of usually
expressed by a relative error value of usually less less than
VCE
than [V]10%. The measurements
10%. The measurements were were carried
carried out to
out to
obtain the transistor
obtain the transistorjunction
junction temperature
temperature (T (Tjj))close
closetotothethe maximum
maximum valuevalue ofparameter
of this this parameter
given ingiven
in thethe
catalogue
catalogueby bythe
themanufacturer.
Figure manufacturer.
5. Non-isothermal The
The junction
junction
output temperature
temperature
characteristics of the waswas estimated
estimated
BT1206-AC on basis
on the
transistor. the basis
of caseof case
temperature
temperature (TC) measurements
(TC ) measurements and and catalog
catalog valuevalue
of theofthermal
the thermal resistance
resistance Rthjc between
Rthjc between the
the transistor
junction As seen,
transistor
and the satisfying
junction
case. and case. agreement between the results of simulations and measurements was
expressed by a relative error value of usually less than 10%.equal
The to
measurements awere
The simulated
The simulated output output characteristic
characteristic at
atthe
the base
basecurrent
current equal 150tomA150has
mA hascarried
specific
a shape. outThe
specific to
shape.
obtain
negativetheslope
transistor
of the junction
collectortemperature
current with(T j) close tocollector-emitter
increasing the maximum value of this
voltage parameter
is caused by thegiven in
strong
The negative slope of the collector current with increasing collector-emitter voltage is caused by the
the catalogue by the manufacturer. The junction temperature was estimated on the basis of case
strong self-heating phenomenon. In the literature, this type of characteristic is called an N-shape
temperature (TC) measurements and catalog value of the thermal resistance Rthjc between the
characteristic
transistor [8].
junction and case.
The Theimportant
simulated parameter in the description
output characteristic at the baseof the bipolar
current equal totransistors
150 mA has aisspecific
the common-emitter
shape. The
current gain factor
negative slope of(β),
the defined as the ratio
collector current of the output
with increasing and controlvoltage
collector-emitter current (β = ICby
is caused /IBthe
). Itstrong
is known
from the literature that the value of β depends on the operating point. In catalogs, usually only one
value of this parameter is given (under set operating conditions). Manufacturers deliberately choose a
specific transistor operating point to present the highest value of the current gain factor β. Figure 6
 Energies 2020, 13,
self-heating x FOR PEER REVIEW
phenomenon. In the literature, this type of characteristic is called an N-shape 7 of 9

characteristic [8].
self-heating
The importantphenomenon.
parameter Ininthe
the literature,
descriptionthis typebipolar
of the of characteristic
transistors isisthecalled an N-shape
common-emitter
characteristic [8].
current gain factor (β), defined as the ratio of the output and control current (β = IC/IB). It is known
Energies The
2020,
from important
13,literature
the 2617 parameter
that the valueinofthe description
β depends of operating
on the the bipolarpoint.
transistors is theusually
In catalogs, common-emitter
only one 7 of 9
current gain factor (β), defined as the ratio of the output and control current (β
value of this parameter is given (under set operating conditions). Manufacturers deliberately choose = IC/IB). It is known

afrom thetransistor
specific literature operating
that the value
pointoftoβpresent
depends theonhighest
the operating
value ofpoint. In catalogs,
the current usually
gain factor only one
β. Figure 6
shows value
the of this
authors’ parameter
measured is given
and (under
calculatedset operating
values conditions).
of β factor Manufacturers
for the deliberately
considered
shows the authors’ measured and calculated values of β factor for the considered transistor in choose
transistorthein the
a specific
isothermal
isothermal transistor
conditions atoperating
conditions aatfixed
a fixed point
valueto
value ofofpresent
the the highest value
the collector-emitter
collector-emitter of the
voltage
voltage current
equal
equal gain
to 3toV.3 V.factor β. Figure 6
shows the authors’ measured and calculated values of β factor for the considered transistor in the
isothermal conditions at60 a fixed value of the collector-emitter voltage equal to 3 V.

50
60 Ta= 25ºC
40
50
β [A/A]
T = 60ºC
Ta= a25ºC

β [A/A] 30
40 Ta= 130ºC
Ta= 60ºC

20
30 Ta= 130ºC

10
20
0
10
0 1 2 3 4 5 6
0 iC [A]
0 1 2 3 4 5 6
Figure 6. Isothermal dependence of the current gain iC [A]
factor at fixed values of an ambient temperature.
Figure 6. Isothermal dependence of the current gain factor at fixed values of an ambient temperature.

As seen, in6.the
AsFigure
seen, in range
Isothermal
the range ofofcollector
dependence current
of the
collector current values
current gain factor
values up
up toabout
about
attofixed values 6ofA,
6 A, an the
β(iCβ(i
theambient C ) dependence
temperature.
) dependence is a is a
monotonically
monotonically increasing
increasing function.
function.In In addition,
addition, thethedependence
dependence of of
thethe current
current gaingain factor
factor β(T)
β(T) at a at a
fixedfixed
value Asofseen,
value aincollector
the range
aofcollector current of collector
current current values
isisa adecreasing
decreasing function.
function. upThisto about
Thisis is a6characteristic
A, the β(iCfeature
a characteristic ) dependence
feature of isbipolar
of bipolar a
monotonically increasing function. In addition, the dependence of the current gain factor β(T) at a
power transistors
power made
transistors madewith silicon
with carbide
silicon carbide technology
technology[8,25–27].
[8,25–27]. The The results
results ofof simulations
simulationsusing using the
fixed value ofβamodel collector current is agreement
a decreasing function. This isofathecharacteristic feature ofthebipolar
proposed β model has a good agreement with the results of the measurements of the β(iCof) characteristics
the proposed has a good with the results measurements β(iC)
power transistors made withof silicon carbide technology [8,25–27]. Thetheresults of simulations using
in allcharacteristics
ranges
the proposed
in all
of temperature,
β model
rangeswhich
has a
temperature,
goodindicates
agreement
which
that with
indicates
the transistor’s
the results
thatcurrent
of the
transistorʹs
gain properties
measurements
currentdeteriorate
of the
gain
β(iC)
properties deteriorate with the temperature increase.
with the temperature
characteristics increase.
Figure 7 presents the isothermal and non-isothermal dependence of the current gain factor βgain
in all ranges of temperature, which indicates that the transistorʹs current of
Figure 7 presents
properties deteriorate thewith
isothermal
the ambient and non-isothermal
temperature increase.at a fixeddependence of the current gain factor β of
the considered transistor on the temperature value of the collector-emitter voltage
the considered
and theFigure transistor
7 presents
control currentthe
onequal
the ambient
isothermal
to 3 V and
temperature
and non-isothermal
50
at a dependence
mA, respectively.
fixedThe value ofthe
of
points
the collector-emitter
current
represent gain factor βvoltage
the results ofof
and thethe considered
control transistor
current equalon the
to 3ambient
V and temperature
50 mA, at a fixed
respectively.
measurements and lines represent the results of calculations using the proposed ETM. value
The of the
points collector-emitter
represent thevoltage
results of
and the control
measurements and linescurrent equal tothe
represent 3V and 50ofmA,
results respectively.
calculations usingThethe points represent
proposed ETM. the results of
measurements and lines 50 represent the results of calculations using the proposed ETM.
VCE= 3 V
50 IB= 50 mA
40 Isothermal VCE= 3 V

40 Isothermal IB= 50 mA
30
β [A/A]

Non-isothermal
30
β [A/A]

20 Non-isothermal

20
10

10
0
20 40 60 80 100 120
0
Ta [oC]
20 40 60 80 100 120
a T [oC]
Figure 7. Isothermal and non-isothermal dependence of the current gain factor.
Figure 7. Isothermal and non-isothermal dependence of the current gain factor.
Figure 7. Isothermal and non-isothermal dependence of the current gain factor.

As seen, the non-isothermal characteristic of SiC BJT lies under the isothermal counterpart.
Moreover, at room temperature the differences between isothermal and non-isothermal measurements
reach over 25%. This phenomenon is due to the fact that when the device lattice temperature and
phonon scattering in SiC BJTs increase, they cause both the carrier mobility and the current gain to
drop [28].
Energies 2020, 13, 2617 8 of 9

4. Conclusions
The authors’ measurements show that the ambient temperature rises strongly affect certain
characteristics and parameters of the silicon carbide power bipolar transistor. Moreover, the self-heating
phenomenon in SiC BJT can result in not only qualitative but also quantitative changing of the
semiconductor device characteristics shape, resulting in the negative slope of collector current with
increasing collector-emitter voltage occurrence and causes decreasing of the current gain factor β with
power dissipation increase.
In the typical Gummel–Poon model implemented in SPICE, the self-heating phenomenon is not
taken into account. In this paper, the authors proposed an improved SiC BJT model dedicated for SPICE,
incorporating not only the self-heating phenomenon, but also modeling the quasi–saturation mode using
the temperature-dependent intrinsic carrier concentration model. Apart from this, the current gain factor
model is used for the first time for the bipolar transistor made of silicon carbide taking the temperature
and collector current impact into account. The proposed electrothermal model of the BT1206-AC
SiC BJT allows calculating the non-isothermal characteristics of the tested device. The simulated and
measured characteristics of considered SiC power BJT were compared. The satisfactory agreement
between both the calculated and measured characteristics indicates the correctness of the presented
model. It is worth mentioning that the accuracy of the proposed model can be achieved even at high
current density when the current gain factor β starts to decrease.

Author Contributions: Investigation, J.P., D.B. and J.Z.; methodology, J.P., D.B. and J.Z.; supervision, J.Z.;
writing—original draft, J.P., D.B. and J.Z.; writing—review and editing, J.P., D.B. and J.Z. All authors have read
and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Wang, J.; Liang, S.; Deng, L.; Yin, X.; Shen, Z.J. An improved SPICE model of SiC BJT incorporating surface
recombination effect. IEEE Trans. Power Electron. 2019, 34, 6794–6802. [CrossRef]
2. Bargieł, K.; Bisewski, D.; Zar˛ebski, J. Modelling of dynamic properties of silicon carbide junction field-effect
transistors (JFETs). Energies 2020, 13, 187. [CrossRef]
3. Zar˛ebski, J.; Dabrowski,
˛ J.; Bisewski, D. Measurements of thermal parameters of silicon carbide semiconductor
devices. Przeglad ˛ Elektrotechniczny 2011, 87, 29–32.
4. DiMarino, C.; Chen, Z.; Boroyevich, D.; Burgos, R.; Mattavelli, P. Characterization and comparison of 1.2 kV
SiC power semiconductor devices. In Proceedings of the 15th European Conference on Power Electronics
and Applications (EPE), Lille, France, 2–6 September 2013; pp. 1–10.
5. Catalog data of SiC BJT BT1206-AC. Available online: https://www.dacpol.eu/pl/elementy-
polprzewodnikowe-z-weglika-krzemu/product/elementy-polprzewodnikowe-z-weglika-krzemu-1224
(accessed on 20 February 2020).
6. Singh, R.; Grummel, B.; Sundaresan, S. Short circuit robustness of 1200 V SiC switches. In Proceedings of the
IEEE 3rd Workshop on Wide Bandgap Power Devices and Applications (WiPDA), Blacksburg, VA, USA,
2–4 November 2015; pp. 1–4.
7. Oueslati, M.; Garrab, H.; Jedidi, A.; Besbes, K. The advantage of Silicon Carbide material in designing of
power Bipolar Junction Transistors. In Proceedings of the 12th International Multi-Conference on Systems,
Signals & Devices, Mahdia, Tunisia, 16–19 March 2015; pp. 1–6.
8. Patrzyk, J.; Zar˛ebski, J.; Bisewski, D. DC characteristics and parameters of silicon carbide high-voltage power
BJTs. In Proceedings of the 39th International Microelectronics and Packaging (IMAPS) Conference, Gdańsk,
Poland, 20–23 September 2015; pp. 1–8.
9. Santi, E.; Peng, K.; Mantooth, H.A.; Hudgins, J.L. Modeling of wide-bandgap power semiconductor
devices—part II. IEEE Trans. Electron. Devices 2015, 62, 434–442. [CrossRef]
10. Górecki, K.; Detka, K. Influence of power losses in the inductor core on characteristics of selected DC-DC
converters. Energies 2019, 12, 1991. [CrossRef]
Energies 2020, 13, 2617 9 of 9

11. Dabrowski,
˛ J.; Krac, E.; Górecki, K. New model of solar cells for SPICE. In Proceedings of the 25th International
Conference Mixed Design of Integrated Circuits and Systems MIXDES, Gdynia, Poland, 21–23 June 2018;
pp. 338–342.
12. Janicki, M.; Torzewicz, T.; Ptak, P.; Raszkowski, T.; Samson, A.; Górecki, K. Parametric compact thermal
models of power LEDs. Energies 2019, 12, 1724. [CrossRef]
13. Mantooth, H.A.; Peng, K.; Santi, E.; Hudgins, J.L. Modeling of wide bandgap power semiconductor
devices-part I. IEEE Trans. Electron. Devices 2015, 62, 423–433. [CrossRef]
14. Tian, Y.; Hedayati, R.; Zetterling, C.-M. SiC BJT Compact DC model with continuous-temperature scalability
from 300 to 773 K. IEEE Trans. Electron. Devices 2017, 64, 3588–3594. [CrossRef]
15. Balachandran, S.; Chow, T.P.; Agarwal, A.; Tipton, W.; Scozzie, S. Gummel-Poon model for 1.8 kV SiC
high-voltage bipolar junction transistor. In Proceedings of the IEEE 35th Annual PESC, Aachen, Germany,
20–25 June 2004; pp. 2994–2998.
16. Johannesson, D.; Nawaz, M. Development of a simple analytical PSpice model for SiC-based BJT power
modules. IEEE Trans. Power Electron. 2016, 31, 4517–4525. [CrossRef]
17. Liu, W. Electro-Thermal Simulations and Measurements of Silicon Carbide Power Transistors. Ph.D. Thesis,
The Royal Institute of Technology, Stockholm, Sweden, 2004.
18. Amin, A.R.; Salehi, A.; Ghezelayagh, M.H. BJT circuits simulation including self-heating effect using FDTD
method. In Proceedings of the Asia-Pacific International Symposium on Electromagnetic Compatibility,
Beijing, China, 12–16 April 2010; pp. 924–927.
19. Cadence Design Systems Inc. PSPICE A/D Reference Guide Version 10.2; Cadence Design Systems Inc.: San
Jose, CA, USA, 2004.
20. Zar˛ebski, J.; Górecki, K. Parameters estimation of the d.c. electrothermal model of the bipolar transistor.
Int. J. Numer. Model. Electron. Netw. Devices Fields 2002, 15, 181–194. [CrossRef]
21. Górecki, K.; Zar˛ebski, J.; Górecki, P.; Ptak, P. Compact thermal models of semiconductor devices: A Review.
Int. J. Electron. Telecommun. 2019, 65, 151–158.
22. Górecki, P.; Górecki, K. Modelling a switching process of IGBTs with influence of temperature taken into
account. Energies 2019, 12, 1894. [CrossRef]
23. Bisewski, D. Parameters estimation of SPICE models for silicon carbide devices. In Proceedings of the
21st European Microelectronics and Packaging Conference (EMPC) and Exhibition, Warszawa, Poland,
10–13 September 2017; pp. 1–5.
24. Feng, Z.; Zhang, X.; Wang, J.; Yu, S. High-efficiency three-level ANPC inverter based on hybrid SiC and Si
devices. Energies 2020, 13, 1159. [CrossRef]
25. Daranagama, T.; Pathirana, V.; Udrea, F.; McMahon, R. Novel 4H-SiC bipolar junction transistor (BJT) with
improved current gain. In Proceedings of the 13th Brazilian Power Electronics Conference and 1st Southern
Power Electronics Conference (COBEP/SPEC), Fortaleza, Brazil, 28 November–2 December 2015; pp. 1–6.
26. Elahipanah, H.; Kargarrazi, S.; Salemi, A.; Östling, M.; Zetterling, C.-M. 500 ◦ C High current 4H-SiC lateral
BJTs for high-temperature integrated circuits. IEEE Electron. Device Lett. 2017, 38, 1429–1432. [CrossRef]
27. Liang, S.; Wang, J.; Peng, Z.; Chen, G.; Yin, X.; Shen, Z.J.; Deng, L. A modified behavior spice model for SiC
BJT. In Proceedings of the IEEE Applied Power Electronics Conference and Exposition (APEC), San Antonio,
TX, USA, 4–8 March 2018; pp. 238–243.
28. Meng, F.; Ma, J.; He, J.; Li, W. Phonon-limited carrier mobility and temperature-dependent scattering
mechanism of 3 C-SiC from first principles. Phys. Rev. B 2019, 99, 045201. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).