CHAPTER 2

REVIEW OF RELATED LITERATURE

This thesis proposal focuses on the indoor light and thermal energy harvesting

technologies. Advancements in each of these areas will be discussed, as well as the

use of more than one technology to create hybrid energy harvesters. Advantages of

using hybrid harvesters with their respective pros and cons of each will be discussed.

Also this chapter also aims to summarize some of the published papers and

references of the RF to DC converter which will the foundation of modifying the

CMOS differential rectifier. These designs include the analysis of various practical

techniques for optimizing performance and tradeoff between parameters.

2.1.1 Indoor Light Energy Harvesting

Many energy harvesting systems have been developed and those reported in the

literature are mostly for outdoor applications. Very few research works have

discussed the indoor energy harvesting from light source, which has serious

challenges to resolve.

With low light intensity, the PV voltage may be too low to charge up an energy

storage/ battery. Thus the power management should cater for the large difference in

light intensities and the voltage step-up scheme is needed.

Traditional DC-DC converter has been used to step up the voltage in recent works

[5]. However, the inductors in the converter are costly if being integrated on chip.

Moreover, most of the strategies tried to maximize the converter power loss, while
the solar power transfer efficiency (measured at the converter output) is the actual

design issue that should be considered.

So [7] presented a micro solar power management system with an inductor

less solution which facilitates the system on chip integration. A charge pump is used

to step up the PV voltage to charge a battery or directly operate the circuit. Shown in

figure 2.1, an inductor-less micro solar power management design for energy

harvesting applications.

Figure 2.1 Block diagram of an inductor-less micro solar power management design
for energy harvesting applications [26].

In research work of [1][11] presents an all on-chip photovoltaic cell energy

harvesting system for low power indoor application with NiMH battery to store the

harvested energy. The system does not use a dc to dc converter in boosting an output

voltage to avoid complicated control algorithm and the costly implementation of

inductor on the chip. Also the circuit is fabricated using 0.18um 1P6M process. [1].

Also [11] presents same architecture but implemented in 90nm CMOS technology.

Shown in figure 2.2 below was the block diagram used by [1] [11].
 Moreover [1][11] uses differential drive rectifier to step up voltage. According

to [1] differential drive rectifier has its good feature same with charge pump. Also it

has the advantage cause it applies active VTH cancellation scheme by using the

cross-couple differential drive configuration to actively biased the transistor

[12].Because of this active biasing, VTH can be minimized in a forward bias condition

which reduces the turn-on voltage and the same time automatically increases in a

reverse bias condition which reduces leakage current.

Shown in figure 2.2 is the block diagram solar energy harvesting circuit of [1]

that uses multi-stage differential rectifier to step up voltage.

Figure 2.2 Block diagram of solar energy harvesting circuit [1].

2.1.2 Thermal Energy Harvesting

A simple TE generator is made by heating one face of TE module, and

cooling the other face causing an electrical current to be generated by connected a

load to the end points of the TE module. This behaviour is described as the Seebeck

effect, and was discovered by Tomas Seebeck in 1821. A TE generator has

demonstrated attractive characteristics such as a long life cycle, no moving parts,

simple and high reliability [17]. However, its low efficiency is a big drawback that

has continually prevented the widespread commercial application of this technology.

Current TE materials can only convert a maximum of 5-6% of the useful heat into

electricity. However, some significant researches is being carried out to develop new

materials and module constructions, which promise harvesting efficiency of more

than 10%.

Thermal energy harvesting can be use of many applications just like, [8]

investigate the feasibility of powering wireless metering devices, namely heat cost

allocators, by thermal energy harvested from radiators. And design a self powered

telemetric device wireless sensor node for temperature measurement. This device is

realized with a conventional off the shelf thermoelectric generator as a power source.

Moreover, there is also some research which focuses on designing a high

efficient TE generator for WSNs. In [9] the authors developed a thermal energy

harvesting system, which can harvest thermal energy from sunlight. In order to

enhance its efficiency, eight TE modules were placed in a small greenhouse. 40mW

power could be generated from such a TE generator when it is placed on a 200

surface. [10] proposed a thermogenerator system to harvest energy from small

temperature gradients between the human body, and the ambient environment. The

power generated by this TE generator is often as low as 0.3V, which is not suitable

for many practical applications. To supply a wireless communication module and to

charge an energy storage element, a charge pump IC with a step up DC-DC converter
was employed to boost the input voltage and etc. Further, Xin Lu et al. [18] presents

a low temperature thermal energy harvesting system, which can harvest heat energy

from a temperature gradient and convert it into electrical energy, which can be used

to power wireless electronics. Specifically it showed to extend the ZBARVs

(ZigBee based radiators valve) lifespan. Base on theoretic calculation, the lifetime of

the whole ZigBee system could be extended for than 8 years.

2.2 Hybrid Energy Harvesting

To overcome the drawbacks of single source energy harvesting and to extend

the lifespan of wireless sensor node, [3], for example, have presented a mechanical

structure made up of two different energy harvesting mechanisms on the same

platform to harvest from the same vibration energy source. Other hybrid energy

harvesting (HEH) approaches discussed in [4][3] focus on different power

management schemes to harvest multiple energy sources.

Guilar et al. [4] and Lhermet et al. [3] proposes to combine energy harvesting

using an electronic switch/ multiplexer to switch between two energy sources.

Whenever both energy sources are present simultaneously, based on the priority given

by the power management circuit, only one of the two energy sources would then be

harvested.

In order to lessen the area of the boost converter Chao Shi et al. [13] proposes

a single inductor multiple input boost converters. Then uses time multiplexed

operation of the proposed converter that enables sharing of the power stage between
different input sources. Shown in figure 2.3 a single-inductor multiple input boost

converter.

Figure 2.3 Single-Inductor multiple input boost converter [13].

Its not always possible to harvest energy from energy souces simultaneously,

so [14] introduces an HEH system, AmbiMax that harvests simultaneously from

multiple energy sources to sustain operation of the sensor nodes. Each energy sources

is allocated to charge its own supercapacitor and energy stored in these supercacitor

is then transferred to the lithium-ion battery. Major drawback as more energy sources

are combined together the number of power management unit for each individual

energy sources increases, hence more components are needed and larger volumetric

size, higher power losses, and higher costs incurred.

To enhance the performance of the wireless sensor node deployed in the

challenging context using HEH from solar and thermal energy sources [2] uses a near

maximum power tracking technique to maximize power transfer from hybrid energy

source to the node. Also it only uses one power management circuit to condition the
combined output power harvested from both sources. An efficient microcontroller

based ultra low power management circuit with fixed voltage reference based

maximum power point tracking is implemented with closed-loop voltage feedback

control to ensure near maximum power transfer.

It can extract maximum output power from hybrid energy harvester with a

simple and ultra low power control circuit so as to place the panel at its MPPs, rather

than using those energy hungry tacking techniques such as Perturb and Observe and

Incremental Conductance, which require high computational power and cost. With an

average electrical power of 621uW harvested at an average indoor irradiance of 1010

lux and a thermal gradient of 10K .

The designed power management circuit with fixed voltage reference MPPT

approach is essentially consist of three main building blocks: 1) a boost converter

with MPPT tracker and its control and its control and pulse width modulation

(PWM) generation circuit that manipulates the operating point of the HEH scheme to

keep harvesting power at near MPPs;2) an energy storage element element i.e.,

supercapacitor to buffer the energy transfer between the source and the load; 3) a

regulating buck converter to provide constant voltage to the wireless sensor node and

other electronic circuitries.

 This hybrid indoor ambient and thermal energy harvesting scheme uses

discrete component. Figure 2.4 shows the functional block diagram of HEH System

[6].

Figure 2.4 Functional block diagram of HEH System [6].

Thus, it is able to harvest simultaneously from both energy sources whenever

they are available instead of harvesting individual source at one time.

2.3 DC Boost Converter, Charge Pump and Differential CMOS Rectifier.

For low power application, switched-capacitor converters such as charge

pump circuit are the primary choice rather than using switching DC converters,

because of acquiring an advantage of lower cost and less area consumption [15].

There are techniques for charge pump architecture that has been introduced to

improve its performance such as using a voltage controller to resolve on the problem

of the charge transfer-switch (CTS) reverse charge sharing effect at each stage and

simultaneously solve the limitation in the threshold voltage by using a cross-coupled

structure on the last stage [15]. The combination of these techniques would

completely compensate all the voltage-drop issues on charge pump structure.

On the other hand, a differential-drive rectifier has its good feature the same

with the charge pump. Differential-drive rectifier is actually a modified charge pump

circuit. As mentioned earlier, the advantage of the differential-drive rectifier from the

other CMOS rectifier architecture is that it applies an active Vth cancellation

scheme by using the cross-couple differential-drive configuration to actively biased

the transistors [12]. Because of this active biasing, Vth can be minimized in a forward

bias condition which reduces the turn-on voltage and at the same time automatically

increases in a reverse bias condition which reduces the reverse leakage current.

Figure 2.5 Charge pump circuit [1].

In [1], charge pump and rectifier are simulated, tested and compared its

performances. Charge pump is composed of three stages as shown in figure 2.5.

Meanwhile, differential-drive rectifier is composed also of three stages as shown in

figure 2.6. The output voltage signal of differential-drive rectifier and the improved
charge pump are illustrated in figure 2.7. Both the structure are tested with input

voltage Vin = 1V, pulse voltage signal of VPULSE = 1V with a frequency of 1MHz,

switch capacitor of 0.1uF and resistor load RLOAD = 10k. Their transistors are set

to be of the same size. The parameters of switched-capacitor structures are listed in

table 2.1 . Both the structures produce an output voltage of 4V at no-load. But charge

pump structure requires more charging time and capacitors due to an extra pumping

stage with cross-coupled transistors at the last stage. Moreover, the rectifier efficient

is very high especially when the voltage amplitude is larger than the threshold

voltage.

Figure 2.6 Multi-stage Differential-drive rectifier [1].

 Figure 2.7 Output voltage signal of differential rectifier and charge pump circuit at
1V input.

Table 2.1 Rectifier and charge pump performance comparison [10].

Differential-drive rectifier Charge Pump [22]
No No
With load With load
load load
10k 1k 10k 1k
Input 1V 1V 500mV 1V 500mV 1V 1V 500mV 1V 500mV
Voltage
Switching 0.1uF
Capacitor
Clock 1MHz
Frequency
Ideal 4V 4V 2V 4V 2V 4V 4V 2V 4V 2V
Output
Voltage
Output 4V 3.99V 1.9V 3.9V 1.16V 4V 3.96V 900mV 3.06V 520mV
Voltage
Number 12 14
of MOS
Number 6 5
of
Capacitor
Pumping 99.85% 95% 96.50% 58% 99.0% 45% 90.00% 26.00%
gain
2.4 RF to DC Converter modified architectures
In the work of [16] , two passive voltage multipliers, namely Modified

Dickson (MDVM) as shown on figure 2.8 and Modified Mandal-Sarpeshkar Voltage

Multipliers as shown in figure 2.9 are designed and implemented. From a mere input

power of 10uW to 60uW at 13.56MHz voltage multipliers effectively increased

output voltage 905mV to 2.128V and 1.114V to 3.609V for MDVM and MSVM,

respectively, given a supply voltage of 1V and capacitive load of 30pF. All designs

are implemented using 90nm CMOS process.

Figure 2.8 Two-stage MDVM schematic diagram [16].

 Figure 2.8 Two-stage MSVM schematic diagram [16].

From the obtained results as shown on Figure 2.9 and Table 2.2, the Modified

Dickson Voltage Multiplier (MDVM) seems to be the better choice when

implementing a passive voltage multiplier because of its lesser number of stages and

smaller area. The choice would generally depend on the desired output voltage. If

VOUT is around 1V, then MDVM is more suitable as it would optimize the area and

number of stages. However, if VOUT is much greater than 1V, say around 3V to 5V

then MSVM is more recommended as it is more linear, is less complex, and

eventually leads to lesser number of stages [16].

Figure 2.9 Comparisons of obtained parameters [16].

Table 2.2 Obtained Results for the Voltage Multiplier [15].

Parameter MDVM MSVM

Input Power 10uW-60uW
Load 30pF
Frequency 13.56MHz
Vout 904.9mV-2.128V 1.114V-3.609V
Current Consumption 225uA 203uA

2.5 Summary
Different indoor light and thermal energy harvesting technologies has been

examined in this literature review. Where in single energy sources might not be

sufficient and efficient sustain the operation or even enhance the performance of

wireless sensor node over lifetime. Thus [2] proposes a readily a hybrid energy

harvesting to overcome the drawback of single energy source.

Also RF to DC converter which will be the foundation of modifying the

CMOS differential rectifier is examined; where in many have done their research to

achieve a higher output voltage by proposing different methods and designs. All [2]

and [10] works are discussing on eliminating the voltage threshold limitation effect of

a transistor by proposing different methods like having an additional charge transfer

block or by designing new converters to improve the charge transfer capability of the

transistor which can be a good method to improve RF to DC rectifier.