2014

Amorphous Silicon: Properties

and Applications

SHANUJ GARG
2014H123162P
11/16/2014

Chapter 1
Introduction
Silicon is plenty available on earth surface in the form of dusts, sands, planetoids, silica etc.
Although most of the silicon is used in the steel refining, aluminium-casting, and fine
chemical industries, a relatively small portion (< 10%) of very high purity silicon is used in
semiconductor industries, which is called Electronic Grade Silicon (EGS).
EGS can be extracted directly from solid silica or other silicon compounds by purification
techniques like float zone refining and fluidized bed (FB) based on melting and resolidifying.
Silicon which we get from these techniques is polycrystalline in nature. From this
polycrystalline silicon Monocrystalline silicon is extracted using CZ growth which is
specifically used in fabrication of integrated circuits.[1]
Apart from these a non-crystalline allotrope of silicon also exists that is called Amorphous
Silicon.[2] It is often abbreviated to a-si.

1.1

Amorphous Silicon: Historical Background

In Dundee, Scotland, Walter Spear and Peter LeComber discovered around 1973 that
amorphous silicon prepared using a glow discharge in silane (SiH4) gas had unusually good
electronic properties; as they were building on earlier work by Chittick, Sterling, and
Alexander. Glow discharges are the basis for the familiar neon light; under certain
conditions, an electric voltage applied across a gas can induce a signicant electrical current
through the gas, and the molecules of the gas often emit light when excited by the current.
Amorphous silicon was deposited as a thin lm on substrates inserted into the silane gas
discharge. Spear and LeComber reported in 1975 that amorphous silicons conductivity could
be increased enormously either by mixing some phosphine (PH3) gas or some diborane (B2H6
) gas with the silane. Just as for crystal silicon, the phosphorus doping of the amorphous
silicon had induced a conductivity associated with mobile electrons (the material was ntype), and the boron doping had induced a conductivity associated with mobile holes (the
material was p-type).[3]
In 1974 David Carlson discovered that he could make fairly efcient solar cells using a silane
glow discharge to deposit lms. In 1976, he and Christopher Wronski reported a solar cell
2014H123162P

Page 1

based on amorphous silicon with a solar conversion efciency of about 2.4%. These two had
discovered, the optoelectronic properties of amorphous silicon made by glow discharge (or
plasma deposition) are very much superior to the amorphous silicon thin lms prepared, for
example, by simply evaporating silicon. After several years of uncertainty, it emerged that
plasma-deposited amorphous silicon contained a signicant percentage of hydrogen atoms
bonded into the amorphous silicon structure and that these hydrogen atoms were essential to
the improvement of the electronic properties of the plasma-deposited material. As a
consequence, the improved form of amorphous silicon has generally been known as
hydrogenated amorphous silicon (or, more briey, a-Si:H) is found. In recent years, many
authors have used the term amorphous silicon to refer to the hydrogenated form, which
acknowledges that the unhydrogenated forms of amorphous silicon are only matter of study
today.[3]
Also amorphous silicon thin film transistors are the key semiconductor of an entire area in
the field of electronics, known as macro-electronics, as crystalline silicon FETs are the
key enablers for the integrated circuit field.[4]

1.2

Requirement of Macroelectronics

Micro-Electronics technology has revolutionized computing and communications associated

with all manner of systems. As device scaling continues, and we move fully into the age
of systems on a chip and systems in a package, these advances will not only continue,
but become more pervasive. Yet, as revolutionary as microelectronics has been, there
are functions that are not well addressed by conventional microelectronics technology. Since
the driving force behind microelectronics has been smaller and smaller devices in smaller
and smaller areas, those applications that require the electronics to be spread over a large
area are difficult or costly to achieve with the conventional approach. For example flat panel
displays where requirement for electronic control at each pixel of the display is there.
Another example is solar cell arrays where photo diodes must be spread over areas as large as
several square meters to collect sunlight sufficient to run the needed equipment. Flexible and
large area electronics using amorphous silicon (a-Si), low temperature polysilicon (LTPS),
and organic & inorganic nanostructured semiconductor materials is a technology that is
beginning to show tremendous promise in these type of cases. Thin film transistors (TFTs)
using a-Si and LTPS on rigid silicon and glass substrates are employed for driving active
matrix liquid crystal displays and image sensors and also in photovoltaics.[5]
2014H123162P

Page 2

Chapter 2
Structure and Properties of a-Si
Till now we have seen the emerging need of large area electronics devices which will be
costlier if fabricated with microelectronic technologies; mainly concern about crystalline
form of silicon. In this chapter we will see the structural and behavioural properties of
amorphous silicon which makes it most important material of latest growing macroelectronic
technologies.

2.1 Atomic structure

a-Si has no long range order like a crystal, in device-grade a-Si most silicon atoms still have
bonds to four neighbouring silicon atoms (Fig. 2.1). Surprisingly, these bonds have a bond
length and an angle between bonds close to those in crystalline silicon, despite the lack of
long range order. The words most and close to are important. About 10% of the silicon
orbitals do not form bonds with other silicon atoms, but nearly all of these form bonds
to atomic hydrogen, which reduces the density of residual dangling bonds and states in
the band gap to ~1016 cm-3 . Because the material contains on the order of ~10% hydrogen,
it is often referred to as hydrogenated amorphous silicon, a-Si:H. Second, because the
bond lengths and angles are not all ideal, many of the bonds are strained.[4]

(a)

(b)

Figure 2.1 (a) schematic representation of atomic structure of, single crystal silicon (left) and
hydrogenated amorphous silicon (right). (b) structure of hydrogenated amorphous silicon.

This whole structure can be understood by making a simple model as described above. If we
build a noncrystalline silicon structure with wooden sticks (to represent covalent bonds) and
wooden balls drilled with four small holes for the sticks (to represent the silicon atoms), we
2014H123162P

Page 3

will have some trouble in making a noncrystalline structure. To avoid a crystalline structure,
we will need to bend the sticks and also the fourth stick cant be connected to anywhere that
will represent the condition of dangling bond. The dangling bond problem is solved by
hydrogen atom which forms a covalent bond with existing vacant bond position, thats how
hydrogenated amorphous silicon forms which is often abbreviated to a-Si:H. Strained bond
is also called weak bond in amorphous silicon and it is generally a result of larger
deviations in bond angle and length. The energy in weak bond is always greater then non
strained bonds thats why the probability of their breaking and making defects in a-Si:H
atomic network is greater.
We should note that in the continuous random network like a-Si:H, the definition of a defect
is modified with respect to the crystal structure. In a crystal any atom that is out of place in a
lattice forms a defect. In the continuous random network an atom cannot be out of place.
Because the only specific structural feature of an atom in the continuous random network is
the coordination to its neighbours, a defect in a-Si:H is the coordination defect. This
happens when an atom has too many or too few bonds. In a-Si:H the defects are mainly
represented by Si atoms that are covalently bonded to only three Si atoms (threefold
coordinated)

and

have

one unpaired electron, a so called dangling bond. Since this

configuration is the dominant defect in a-Si:H, the defects in a-Si:H are often related to the
dangling bonds. Also because of hydrogen a significant percentage of dangling bond is been
passivated and this hydrogen passivation of the dangling bonds is responsible

for

semiconductor properties of amorphous silicon.

2.2 Density of States

In a-Si:H, there is a continuous distribution of density of states and no well-defined band gap
exists between the valence band and the conduction band. Due to the long range order
disorder in the atomic structure of a-Si:H the energy states of the valence band and the
conduction bands spread into the band gap and form regions that are called band tail states.
The band tail states represent the energy states of electrons that form the strained bonds in the
a-Si:H network. The width of the band tails is a measure for the amount of disorder in a-Si:H
material. More disorder in a-Si:H means that the band tails are broader. In addition, the
dangling bonds introduce allowed energy states that are located in the central region between
the valence band and conduction band states.[6]

2014H123162P

Page 4

The electron and hole wavefunctions that extend over the whole structure are characteristic
for energy states, in which the charge carriers can be considered as free carriers. These states
are non-localised and are called extended states. The wavefunctions of the tail and defect
states are localized within the structure and therefore these states are called localised states.
Consequently, mobility that characterises transport of carriers through the localised states is
strongly reduced. This feature of a sharp drop of mobility of carriers in the localised states in
a-Si:H is used to define its band gap. This band gap is denoted by the term mobility gap, E
mob, because the presence of a considerable density of states in this gap. Conflicts the
classical concept of the band gap. The energy levels that separate the extended states from the
localised states in a-Si:H are called the valence band and the conduction band mobility edges.
The mobility gap of a-Si:H is larger than the band gap of single crystal silicon and has a
typical value of 1.8 eV. The localised tail and dangling bond states have a large effect on the
electronic properties of a-Si:H. The tail states act as trapping centres and build up a space
charge in a device, the dangling bond states act as very effective recombination centres and
affect in particular the lifetime of the charge carries. The Density of States of a-Si:H is been
shown in the figure below.[6]

Figure 2.2 The schematic representation of the distribution of density of allowed energy states for electrons for
a-Si:H.

Now after this much study about a-Si:H we can put these important differences between
Crystalline and amorphous silicon.
Crystalline Silicon

Amorphous Silicon

1.1ev indirect band gap

1.7-1.9ev direct band gap

4 fold co-ordination

4 fold co-ordination; 3 fold commonly

Fixed bond lengths and angles

Variable bond lengths and angles

2014H123162P

Page 5

Long range order

Very short range order

Higher mobility of carriers ~1000.

Lower mobility of carriers ~1-10 .

We are now able to understand the following properties of a-Si:H.

2.3 Optical Properties

The optical properties of a-Si:H are usually characterised by its absorption coefficient and a
value of the optical band gap. The figure 2.3 shows that a-Si:H absorbs almost 100 times
more than c-Si (crystalline Si) in the visible part of the solar spectrum. The higher absorption
is due to the disorder in the atomic structure of a-Si:H that behaves like a direct gap
semiconductor. This means that 1 m thick a-Si:H layer is sufficient to absorb 90% of usable
solar light energy. In practice the thickness of a-Si:H solar cells is less than 0.5 m that is
about 100 times less than the thickness of a typical single crystal silicon cell. This results in
important savings in both material and energy in fabrication of a-Si:H solar cells.

Figure 2.3 Absorption coefficient as function of photon energy for a-Si:H[6]

Another advantage of a-Si:H is that the optical absorption can be slightly changed by
varying its hydrogen content, and it can be greatly changed by alloying with carbon
2014H123162P

Page 6

or germanium. The absorption coefficient of hydrogenated amorphous silicon carbide (aSiC:H) and hydrogenated amorphous silicon germanium (a-SiGe:H) that are fabricated
at Delft University of Technology are shown along with absorption coefficient of a-Si:H in
Figure 2.3. This feature of easy alloying of a-Si:H allows to design solar cell structures in
which a-Si:H based materials with different absorption properties can be used as active
layers.[6]

2.4 Electrical properties

The electrical properties of a-Si:H are usually characterised in terms of dark conductivity and
photoconductivity. The measurement of these two properties is a standard approach to obtain
information about the quality of a-Si:H material for application in solar cells. Also it gives
information about the mobility-lifetime product and the influence of impurities in a-Si:H. The
mobilities of the charge carriers in the extended states of a-Si:H are about two orders of
magnitude lower than in single crystal silicon. Typically, the electron mobility is 10 cm2 /Vs,
and the hole mobility is 1 cm2 /Vs in intrinsic a-Si:H. The low values of electron and hole
mobilities and the high mobility gap of a-Si:H result in a low dark conductivity, which in
device quality intrinsic a-Si:H is less than 110-10-1cm-1.[6]

2.5 Some other properties

a-Si:H layers can be made thinner than c-Si, which may produce savings on silicon
material cost.

a-Si:H can be deposited at very low temperatures, e.g., as low as 75 degrees Celsius.
This allows for deposition on not only glass, but plastic as well, making it a candidate
for a roll-to-roll processing technique.

Once deposited, a-Si can be doped in a fashion similar to c-Si, to form p-type or ntype layers and ultimately to form electronic devices.

a-Si can be deposited over large areas by PECVD. The design of the PECVD system
has great impact on the production cost of such panel therefore most equipment
suppliers put their focus on the design of PECVD for higher throughput, that leads to
lower manufacturing cost. (PECVD is a a-Si:H deposition technique which is
discussed in the appendix part of this report)

2014H123162P

Page 7

Chapter 3
Applications of a-Si:H
Based on the properties discussed above we can now understand the basic principle of
working of devices fabricated with amorphous silicon. In now a days amorphous silicn has
become a material of choice for TFTs and also solar cells. Basics of these applications are
discussed.

3.1

Amorphous Silicon based Thin Film Transistor Display

Amorphous silicon has become the material of choice for the active layer in thin-lm
transistors (TFTs), which are most widely used in large-area electronics applications, mainly
for liquid-crystal displays (LCDs). LCDs based on TFT are called TFT-LCD which are active
matrix displays contrasting to the previous passive matrix displays. Passive matrix displays
are flat-panel display consisting of a grid of horizontal and vertical wires. At the intersection
of each grid is an LCD element which constitutes a single pixel, either letting light through or
blocking it.[7] When a voltage is applied between two points one from each horizontal and
vertical wire, the corresponding crystal realigns and changes the light transmission properties
of the liquid crystal. By repeating this process, scanning through the pixels an image can be
formed on the display.[8]

Figure 3.1 passive matrix addressing[8]

Problems arise as the number of rows and columns increase. With higher pixel
density, the electrode size must be reduced and the amount of voltage necessary to drive the
2014H123162P

Page 8

display rapidly increases. The higher driving voltage creates a secondary problem; charging
effects. Even though only one row and column are selected, the liquid crystal material near
the row and column being charged are affected by the pulse. The net result is the pixel
selected is active (dark), but the areas surrounding the addressed point are also partially
active (greys). The partially active pixels reduce the display contrast and degrade image
quality. This effect is known as cross talk. Active matrix amorphous silicon TFT displays are
used to overcome this problem.[8]
Although AMLCD (Active Matrix Liquid Crystal Display) technology is more expensive but
solves the scanning limitations of passive displays, contrast ratio and grey scale, and the
flicker induced by holding signals in pixels in a previous LCDs.

Figure 3.2 Active Matrix liquid crystal display[8]

The diagram above (Figure 3.2) shows how thin film transistors are incorporated into a
matrix isolating the pixels. The gates are connected together in rows called gate lines or
buses, and the data lines are connected to the transistor sources. The liquid crystal is
represented as a capacitor. pixel isolation eliminates crosstalk. Also isolation from the
column line permits the pixel capacitor to remain charged, so that faster responding liquid
crystals can be used. Because of these advantages larger displays can be realized.[8]
The dominant active matrix technology is thin-film transistors (TFTs) of a-Si:H. The lower
process temperature (<400 degrees centigrade) of a-Si has led to its dominance in large-size
AMLCD.

2014H123162P

Page 9

3.2

Amorphous Silicon based TFT X Ray Sensor

Large format digital x-ray image sensors are a recent development in the fields of medical
imaging and non-destructive testing. Such image sensors have become practical through the
emergence of large-area, amorphous Silicon (a-Si) TFT and photodiode technologies.[9]
X-ray imaging has been used in many ways, including high resolution imaging using
film, real-time video imaging and digital imaging for digital subtraction angiography
and computer-aided tomography. Additional uses of digital x-ray imaging include bone
mineral densitometry, portal imaging for radiotherapy, and many areas of materials
monitoring which use x-rays for non-destructive testing.[9]
In many of these applications, a large-area flat-panel imager based on a-Si thin-film
transistor (TFT) technology is an attractive component due to its light weight and small
form-factor, high photosensitivity, and lack of image distortion that is present in the
previously used X-Ray techniques.[9]
Many of these applications can be met using a large-area array of a-Si TFTs and photodiodes.
The TFTs are used as pixel switches which address each row of the array, and photodiodes at
each pixel location convert incident light to charge, which is read out by charge amplifiers
connected to each column of the array. The array is patterned on a large glass substrate,
using technology similar to that used in fabricating active-matrix, liquid crystal displays.
Gate driver and charge amplifier ICs are connected to each line on the glass.[9] The difference
between TFT based X-Ray sensor and the old X-Ray is shown by the figure below.

Figure 3.3 difference between the images of old X-Ray sensor (left) and a-Si:H TFT based sensor(right)

3.3

a-Si:H based Solar Cells

As we have seen in previous discussions that by desired doping we can make n type and p
type hydrogenated amorphous silicon, this kind of fabrication is used in a-Si:H based solar

2014H123162P

Page 10

cells. With manufacturing advantages such as scalability and cross-compatibility with the flat
panel display industry, and with considerations for potential scarcity of silicon, the
amorphous silicon (a-Si) thin film photovoltaic devices (TFPD) are often the technology of
choice

for

high

volume

solar

cell

manufacturers.

For understanding the advantage of a-Si:H based solar cells we have to compare it with
conventional

solar

cells

based

on

single

crystal

silicon

structure.

There is a fundamental difference between c-Si and a-Si:H solar cells. A typical c-Si
solar cell, shown in Figure 3.4, consists of a p-type wafer, which is 300 to 500 m thick, with
a thin n-type layer on the top of the wafer, which forms a p-n junction. The width of the
depletion region of the p-n junction is less than 0.5 m. The depletion region, where
an internal electric field is created, represents only a tiny part of the wafer. Most electronhole pairs are generated in the bulk of the electrically neutral p-type region. Electrons, which
are the minority carriers in the p-type region, diffuse towards the p-n junction and in the
depletion region of the junction the electrons drift to the n-type layer under the influence of
the internal electric field. This solar cell is called a diffusion device because the
dominant transport mechanism of the photo-generated carriers is diffusion. Therefore,
the diffusion length of minority carriers (i.e. electrons in p-type material) is an
important material parameter that determines design and performance of c-Si solar cells.
Typical values of the diffusion length of electrons in the p-type float zone single crystal
silicon wafers are in the range of 250 m to 700 m. These values correspond to wafers with
a resistivity in the range of 0.25 to 1.0 cm, respectively.[6]

Figure 3.4 solar cell structure based on c-Si (left) and a-Si:H (right)[6]

2014H123162P

Page 11

In a-Si:H the diffusion length of the charge carriers is much shorter than in single
crystal silicon. In device quality intrinsic a-Si:H the ambipolar diffusion length ranges from
0.1 to 0.3 m. In doped a-Si:H layers, in which the defect density due to the doping is two or
three orders of magnitude higher than in the intrinsic a-Si:H, the diffusion length of
the minority carriers is much lower. A solar cell structure based on the diffusion of the
minority carriers in the electrically neutral regions of the doped layers as in the case of c-Si
does not work in the case of a-Si:H. The photo-generated carriers would all recombine in the
doped layers before reaching the depletion region of the p-n junction due to the very short
diffusion length. Therefore, an a-Si:H solar cell is designed differently. A schematic lay-out
of an a-Si:H solar cell is shown in Figure 3.4. The active device consists of three layers: a ptype a-SiC:H layer, an intrinsic a-Si:H layer, and an n-type a-Si:H layer. The doped
layers are usually very thin, a p-type a-SiC:H layer is ~10 nm thick and an n-type a-Si:H is
~20 nm thick. The active layer in the a-Si:H solar cell is the intrinsic layer which is typically
300 to 500 nm thick. This intrinsic layer is sandwiched in between the doped layers so that
an internal electric field is present across the intrinsic layer. The electron-hole pairs that
are generated in the intrinsic a-Si:H layer immediately experience the internal electric field
that separates electrons and holes from each other. The separated carriers drift under the
influence of the internal electric field towards the doped layers (electrons towards the n-type
layer and holes towards the p-type layer) and are collected by the electrodes. The dominant
transport mechanism of the photo-generated carriers is drift in the internal electric field and
therefore an a-Si:H solar cell is called a drift device. A stack, which consists of p-type
a-SiC:H, intrinsic a-Si:H and n-type a-Si:H layers, forms the p-i-n junction. This structure of
a-Si:H solar cell is presented in Figure 3.4 and it is generally called a single junction a-Si:H
solar cell. The band diagrams of c-Si and single junction a-Si:H solar cells are presented in
Figure 3.4, which also demonstrate the difference between the diffusion type and the drift
type solar cell.[6]
The thickness of the a-Si:H solar cell is comparable to the thickness of the depletion region
in the crystalline silicon solar cell, i.e. about 0.5 m. As mentioned earlier, this small
thickness is sufficient for absorption of the visible part of solar radiation. The small thickness
implies a large reduction in material and energy consumption during production when
compared to crystalline silicon solar cells. Furthermore, when deposited on a light substrate
such as a flexible foil the weight of a-Si:H solar modules is strongly decreased. This

2014H123162P

Page 12

is desired specifically in space applications. The thicknesses of individual layers in c-Si and
a-Si:H solar cells are compared in Table below.[6]
Crystalline silicon solar cell

a-Si:H solar cell

Al contact

0.5 to 1 m

Top glass

SiO2 or Si3N4 layer

~0.1 m

Transparent electrode 0.5 to 0.9 m

n+ type emitter

0.3 to 0.5 m

p type a-SiC:H

0.01 m

p type base

300 to 500 m

Intrinsic a-Si:H

0.3 to 0.5 m

p++ type region

1.0 m

n type a-Si:H

0.02 m

SiO2 layer

0.05 to 0.1 m

ZnO layer

~0.1 m

Al contact

0.5 to 1.0 m

Al or Ag contact

0.5 to 1.0 m

2 to 4 mm

Apart from these potential applications in recent time new applications for a-Si cells are
being sought such as building-integrated PV, space power, consumer electronics, grid
integration, and large scale power generation.

2014H123162P

Page 13

Chapter 4
Conclusion
Amorphous silicon has become the material of choice for large area electronics, with major
applications in displays, image sensing and solar power conversion. The technology
development has occurred because amorphous silicon is a thin film semiconductor that can be
deposited on large, low cost substrates. Furthermore, it retains the material properties of
silicon, and can therefore be patterned into complex electronic devices using the same
photolithographic techniques that have proved so powerful for crystalline silicon electronics.
The report describes the devices, arrays and applications enabled by amorphous silicon, and
how the unique properties of the material determines the design and performance of the
various applications. Active-matrix liquid crystal displays for high performance laptop
computers are a major business, and have established the array manufacturing technology.
Digital X-ray image sensors are poised to revolutionize medical diagnosis by replacing film
with flat-plate amorphous silicon arrays. Solar cells are the most mature application of
amorphous silicon, and many innovations like multi junction solar cell, to enhance the cell
efficiency have been introduced.

2014H123162P

Page 14

Appendix
Plasma Enhanced Chemical Vapour Deposition (PECVD)
The most commonly used deposition method to produce "device quality" a-Si:H, both on the
laboratory

and

industrial

scale,

is

the

radio

frequency

(13.56

MHz)

plasma

decomposition of SiH4 , known as the Plasma Enhanced CVD (rf PECVD) method or the
glow discharge (GD) deposition. The PECVD deposition system is schematically shown in
Figure below. The role of the plasma is to provide a source of energy to dissociate the SiH4
molecules. This is done by collisions with electrons, which originate as secondary electrons
in the plasma and build up their energy by acceleration in an electric field. The growth of an
a-Si:H film is accomplished by attachment of reactive particles of dissociated SiH 4
molecules, called radicals, to the surface of the growing film. As the thickness of the a-Si:H
film for device applications is around half a micrometer, a-Si:H must be deposited on
an appropriate substrate carrier. Some of the energy transferred to the SiH 4 molecules in
the collisions with electrons is radiated as visible light, hence the deposition method is
also called the glow discharge.

Figure PECVD schematic view

2014H123162P

Page 15

References
1. About Silicon (http://www.addisonengineering.com/about-silicon.html) Dated October
10, 2014
2. Amorphous silicon (http://en.wikipedia.org/wiki/amorphous_silicon) Dated October
10, 2014
3. Xunming Deng and Eric A. Schiff Amorphous Siliconbased Solar Cells University
of Toledo, Toledo, OH, USA, Syracuse University, Syracuse, NY, USA.
4. J.C. Sturm, Y. Huang, L. Han, T. Liu, and S. Wagner Amorphous Silicon: The other
Silicon 12th International Conference on Ultimate Integration on Silicon (ULIS), IEEE
(2011)
5. Chalamala B.R., Moussessian A., Kane, Michael G. Macroelectronics: Perspectives
on Technology and Applications Proceedings of the IEEE (Volume:93 , Issue: 7 )
6. Arvind Shah Thin-Film Silicon Solar Cells Chapter 7 (Page 7.3-7.5)
7. Passive-Matrix Displays
(http://www.webopedia.com/TERM/P/passive_matrix_display.html) Dated October 29,
2014
8. Liquid Crystal Display (LCD) Passive Matrix and Active Matrix Addressing
Hitachi application note[AN-002].
9. Richard L. Weisfield Amorphous Silicon TFT X-Ray Image Sensors International
Electronics Devices meeting (1998)

2014H123162P

Page 16