s***l
发帖数: 2236 | 1 I find this could be an interesting topic to talk about.
The competition between Si and other materials, particularly
III-V compounds is really interesting.
When III-V compounds first come to people's mind, lots of them
think they will substitute Si since their characteristics
are much better than Si in many ways, such as high freq
response, high gain, low loss, etc.
The most important drawback of III-V compounds is fabrication.
And that's just the advantage of Si based device.
So with time goes | t*****m
发帖数: 254 | 2 GaAs dominates the microwave circuit and devices in these years.
Because Si technology is well developed and fabrication cost is low, many
people are still trying to extend silicon circuit to higher frequency range.
Though MMIC is fabricated mostly with GaAs, some silicon based MMIC designs
at S or L band are coming up. But they are still in lab. The atracting thing
of these silicon based designs is their lumped element design at this
frequency. This means they can be built much smaller than dis
【在 s***l 的大作中提到】
: I find this could be an interesting topic to talk about.
: The competition between Si and other materials, particularly
: III-V compounds is really interesting.
: When III-V compounds first come to people's mind, lots of them
: think they will substitute Si since their characteristics
: are much better than Si in many ways, such as high freq
: response, high gain, low loss, etc.
: The most important drawback of III-V compounds is fabrication.
: And that's just the advantage of Si based device.
: So with time goes
| m*****a
发帖数: 67 | 3 多孔硅是硅经化学或电化学腐蚀而得(也有人'长'多孔硅). 九十年代
初发现相当强的光致发光(不是荧光)(PL)既而电致发光(EL). 但机理
至今不明. 流行的说法是quantum confinement, 也有人认为是某种
表面效应, 等等. 有兴趣可检索'porous silicon' and 'Canhan'(first
guy), 'Fauchet, p'(my boss) or 'Sailer, m' 等人的文章.
最初几年很热. 然而最近已然冷却. 由余可能是本质的问题至今没有
做成高质量传统器件. 近来的热点是sensor, 尤其是biomedical sensor.
偶的officemates们都干这个(ppmm呀!!). | b**d
发帖数: 7 | 4
No, the two fabrication precesses are quite different.
GaAs for example has no good passivation layer unlike SiO2 for Si.
So MOS structure is difficult to achieve on GaAs.
For III-V integrate circuit, MESFET is dominant.
In addition, III-V materials are less thermally-stable than Si, this also
lead to quite processing techinque.
by the way, I have to mention compound are not alloy as you said, although
GaAs is relatively more difficult than Si to get, now good quality GaAs
crystal are available | s*******r
发帖数: 921 | 5 As far as I know, the cost of GaAs wafers are much more
than silicon wafer, which makes them less widely used even
that they have higher electron mobility.
Also, at very small dimension, the mobility won't be the
limiting factor
as velocity saturation will play the limiting role. At this
condition,
the advantage of III-V over Si will be very small.
Another reason making III-V less attractive in ULSI is that
SiGe is attracting
much attention from the industry, especially IBM. Ge has
also a high m
【在 b**d 的大作中提到】
:
: No, the two fabrication precesses are quite different.
: GaAs for example has no good passivation layer unlike SiO2 for Si.
: So MOS structure is difficult to achieve on GaAs.
: For III-V integrate circuit, MESFET is dominant.
: In addition, III-V materials are less thermally-stable than Si, this also
: lead to quite processing techinque.
: by the way, I have to mention compound are not alloy as you said, although
: GaAs is relatively more difficult than Si to get, now good quality GaAs
: crystal are available
| b**d
发帖数: 7 | 6
No, GaAs wafer is not so expensive compared with SiC, for example.
, usually 3 inch with decent quality costs about 3-5 hundreads bucks.actually
quite cheap.In addition, it is worth noting that the cost of substrate is
quite small in the total of chip fabrication. <2%? The main obstable is its
sophisticated processing. for example, usually 1/3 cost of fabrication is
lithography.
No, the saturation velocity of electron in GaAs is much higher than in Si.
in adiition, at sub micro devide, what dom
【在 s*******r 的大作中提到】
: As far as I know, the cost of GaAs wafers are much more
: than silicon wafer, which makes them less widely used even
: that they have higher electron mobility.
: Also, at very small dimension, the mobility won't be the
: limiting factor
: as velocity saturation will play the limiting role. At this
: condition,
: the advantage of III-V over Si will be very small.
: Another reason making III-V less attractive in ULSI is that
: SiGe is attracting
| s*******r
发帖数: 921 | 7
【在 b**d 的大作中提到】
:
: No, GaAs wafer is not so expensive compared with SiC, for example.
: , usually 3 inch with decent quality costs about 3-5 hundreads bucks.actually
: quite cheap.In addition, it is worth noting that the cost of substrate is
: quite small in the total of chip fabrication. <2%? The main obstable is its
: sophisticated processing. for example, usually 1/3 cost of fabrication is
: lithography.
: No, the saturation velocity of electron in GaAs is much higher than in Si.
: in adiition, at sub micro devide, what dom
| h****i
发帖数: 302 | 8
CMP or post-packaging belongs to processing.
If only for the cost of substrate, it is really very small.
You are both not right.
Actually the satuation velocity of GaAs is only several times lager than Si.
The fact is the critical field for the electron to reach saturation velocity
in the two materials is quite different. For GaAs, about 4kV/cm, but for Si,
it needs >20KV/cm.
The package density of GaAs devices is quite low compared with Si. he means
MMIC.
【在 s*******r 的大作中提到】
|
|
