Patent Application
20140097426
Field effect transistors are known in which source and drain electrodes are bridged by a semiconductor the conductivity of which is controlled by a gate electrode insulated from the semiconductor and the source and drain electrodes to which a potential is applied for this purpose.
In order to meet performance specifications it is necessary for such devices to be acceptably consistent in their performance. This is particularly important in display devices containing many transistors in which inconsistent performance can show as defects in the display.
US 2010/0155710 discloses the process of forming well oriented crystals by isotropic growth and the large crystals of the semiconductor in the channel between the source and drain electrodes appear in the photographs to bridge the channel. This is achieved by depositing a drop of water on its drain electrode and depositing an immiscible solution of a semiconductor on to the drop of water which solution flows over the water droplet into a surrounding channel between drain electrode and the source electrode. The channel seems necessarily to surround the drain electrode but this is difficult to reconcile with some of the drawings and with the photograph of FIG. 9 in which the source and drain electrodes are said to be square shaped.
The procedure seems to have been carried out on a small scale using pipettes or the like and the channels were 200 μm long in the direction of current flow. There are difficulties in scale up to industrial production and in dealing with the short channel lengths of as little as 4 μm now being adopted.
Problems mentioned as requiring attention in the specification include liquid amounts, liquid viscosity, liquid evaporation rates, drop height, drop angle, drop atmosphere, drop splash etc. The problem of delivering droplets onto very tiny targets and then delivering second droplets onto the top of the first droplet are believed to be beyond the practicality of presently known equipment. The problem of preventing variations in the size of the droplets impacting the target and other difficulties are likely to introduce non-uniformity in transistors. The process described appears to be of great difficulty and there is a need for a simpler process of producing transistors of closely reproducible performance on a large scale.
A further problem with the products of this process is that a substantial part of the drain electrode would not be covered by the semiconductor although the degree to which this would be the case would presumably depend on the extent to which the impact of the semiconductor solution splashed or spread the water over the surface. If the water spread into the channel as a result the semiconductor would presumably not fill the channel and its evaporation might damage the connection.
This invention enables the large-scale production of transistors of satisfactorily uniformity of properties by a simpler procedure using semiconductors in which the crystallites do not have to be well oriented.
In WO 03/058729 semiconductor devices are described in which the channel separating the source and drain electrodes is extended by adopting what is described as a non-linear formation. In most of the drawings the non-linear channels are composed of connected linear sections and no curvature is present. Such an arrangement can achieve close packing in which the width of the channel can be maximised. In accordance with the convention used in this art we define channel length as being the distance between source and drain electrodes and channel width to be the geometrical extent of the said channel.
There is no teaching relating to the crystalline nature of the semiconductors or uniformity of performance of transistors, in this reference but in order to emphasise the generality the invention in
If the semiconductor is amorphous it will tend to be consistent with any orientation of the source and drain electrodes, but higher charge mobilities are being sought by the use of crystalline semiconductors which are in general not isotropic in their electrical characteristics.
Consequently if the semiconductor crystals are aligned in an optimum orientation relative to the path of current between the electrodes their performance is better than if their alignment is not optimal. In normal manufacturing procedures the direction of alignment cannot be controlled and the resulting transistors may be of unpredictable performance when manufactured by the same technique. For crystals much smaller than the device channel length, if the orientation of the crystals relative to the path of the current is completely random although the performance of the device will not be optimum each example of the device will have a similar performance. In view of the constant pressure to reduce the size of transistors one will nonetheless find that as the numbers of crystals relative to the channel length declines then localised areas of different average alignment will arise.
Insofar as the crystalline semiconducting material is a single crystal it is fully aligned but not necessarily in the best orientation. If it comprises a large number of crystals it is likely that the crystals will be substantially random in their orientation. If this is not so the average orientation will differ from one device to another and if as is normal the conductivity of the crystals is non-isotropic the performance of the devices will be non-uniform. The shorter the channel length the fewer crystals will bridge it (assuming similar crystal size) and the greater the variation in charge mobility is likely to be.
The semiconductor may be organic or inorganic. It may comprise a single crystalline component which may be a small molecule component or may be a polymeric component which has crystalline and optionally amorphous regions. It may comprise both a crystalline small molecule semiconductor and a binder which may also be a semiconductor. In the latter case crystallites of the small molecule component may be present in a matrix of the binder or may to a greater or lesser extent segregate as a film in the conducting channel thus connecting the electrodes.
If the semiconductor is deposited from solution followed by evaporation of the solvent, crystals formed by the small molecules may be differently aligned by crystal growth phenomena in different channels. If crystal forming semiconducting small molecules are deposited by vapour deposition different alignment in different channels may also be brought about by crystal growth phenomena. If crystalline material is present as particles dispersed in a matrix and the ratio of the average crystal size to the length of the conducting channel is low and there are a large number of crystals present their orientation may approach randomness but if the average crystal size relative to the length of the conducting channel is high for example greater than one quarter, i.e. a ratio of 1 or more to 4 (expressed as the shortest distance from the source to the drain electrode) the uniformity of alignment is more likely to vary from location to location.
There is a trend to reduce the dimensions of field effect transistors and this necessarily brings the source and drain electrodes closer together. As fewer crystallites bridge the gap a variation of performance between different transistors can arise as some regions of crystallites will conform more to the optimum orientation for current flow than others.
This invention comprises a field effect transistor which comprises source and drain electrodes which are bridged by a semiconductor which comprises semiconducting crystallites of varying alignment, the conductivity of the semiconductor being controlled by a gate electrode, which is insulated from the semiconductor and the source and drain electrodes, to which a potential is applied for controlling the conductivity of the semiconductor, in which at least part of the facing surfaces of the source and drain electrodes are geometrically formed such that they provide current flow of different directions between the electrodes through the said channel. By this means current is caused to flow through more orientations of the crystals resulting in greater uniformity of performance between different transistors.
It will be realised that the invention produces the greatest benefit in improving the uniformity of performance of field effect transistors in which the crystallites tend to have a high degree of varying crystallographic alignment, but benefit may also be obtained to a lesser extent when the degree of crystallographic alignment is less.
The invention also provides a method of improving the uniformity of performance of field effect transistors which comprise semiconductors which comprise semiconducting crystallites bridging their source and drain electrodes by the use of source and drain electrodes of which the facing surfaces are contoured so as to have variable tangents.
It is preferred that the electrodes' facing surfaces should be complimentary so that the channel between them is similar in width throughout. The channel is preferably curved throughout its width. If the channel is circular the current will flow in different directions as the inner electrode will have a shorter circumference than the outer one and the current will flow in a fan like pattern. However a circular channel may cause manufacturing difficulties in that the inner electrode has to be accessed through a via and as this can complicate the manufacturing process we prefer open channels, for example a split circle formation in which alternately inverted semicircles are present or by a sinusoidal or similar form, for example connected and alternately inverted sections of the of conics, for example sections of parabolas or ellipses.
The channel should preferably be smooth and substantially free from discontinuities. The facing surfaces are preferably of matching conformation. The length of the channel between them is preferably substantially constant. It is preferred that the geometrical environment of the semiconductor should vary along the whole width of the channel.
The invention is of increasing advantage as the separation of the source and drain electrodes is reduced and it is very useful if the separation is less than 20 microns. It is particularly useful if the separation is at most 10 microns, for example at most 4 microns.
The invention is particularly suitable if the individual crystallites are large enough to be capable of bridging the gap between the electrodes or are approaching such size.
It is preferred that the semiconductor comprises a binder which is preferably amorphous and which is itself preferably a semiconductor. Both crystalline and amorphous semiconductors are well established in the art and a combination of a crystalline semiconductor, for example a polyacene, for example a pentacene which is preferably substituted by trialkylsilyl ethynyl groups and an amorphous semiconductor for example a polytriarylamine is suitable. The invention is believed however to be of a fundamentally physical character related to the orientation of semiconducting crystals as before discussed rather than being related to any particular crystals or any particular location thereof, whether for example present as a film bridging the channel or preferably as a matrix.
One form of the invention will now be described with reference to the drawings in which the source and drain electrodes are identified as 01 and the gate electrode is shown as 02. The shaped transistor source, drain and gate electrodes are shown in bottom gate configuration to show the channel shape in detail. The preparation of the transistors and their performance is described in the experimental results below.
To investigate the effect of the Source and Drain (SD) electrode shape on the spread of drive currents in Organic Thin Film Transistors (OTFT), top gate bottom contact transistors were fabricated. In the experiments a range of different 4″ square substrates were used. These were glass, planarised polyethylenenaphthalate (PEN) (ex Du Pont Teijin Films), and glass coated with a layer of crosslinked SU8 negative photoresist (Microchem corp). Au source and drain electrodes were defined upon the substrate using standard photolithography and chemical wet etching. For the PEN and glass substrates a thin (5 nm) Titanium adhesion layer was used underneath the Au. For the SU8 on glass substrate, the Ti adhesion layer was unnecessary since Au adheres well enough to the SU8 surface throughout the patterning process. The patterns used to define the source and drain electrodes can be seen in
Electrical characterization was performed using a Keithley 4200 semiconductor parameter analyser connected to a Wentworth semi-automatic probe station. For each formulation and electrode shape tested, a minimum of 50 transistors were measured at different points across the substrate. All OTFT were tested using a gate voltage, VG range of +20V to −40 V, while applying a constant drain voltage, VDS of −5 V (source grounded) and measuring the corresponding drain current, Is transfer curve. The linear mobility, μlin of each transistor was calculated according the equation:
where L is the transistor length, W is the transistor width and Ci is the dielectric capacitance per unit area. On current was defined as the current measured for each transistor at a value of Vgs=−40V and Vds=−5V. To exclude the small proportion of devices with gate leakage, a ratio of the gate current to the source-drain current was made at the voltage of Vg −40V, Vds −5V. If this ratio was below 10 (i.e. the gate current was more than 10% of the source drain current, then the device was excluded from the results). Table 1 shows a summary of the main results obtained on 3 substrate types, using formulations at various PTAA:TIPS mixing ratios and with either linear or wavy electrode shape. A 4wt% formulation of PTAA in tetralin was used as to demonstrate that shaped electrodes are only effective where the OSC layer is crystalline (PTAA is amorphous therefore unaffected by the channel orientation). In all cases where TIPS is used the standard deviation of the on current is substantially smaller when the wavy transistor shape is used. In many cases the standard deviation is less than half that of the linear shaped transistor. For the amorphous PTAA material the variation in on current is equal in both wavy and linear transistors (to within experimental error). The best Ion variation for the TIPS/PTAA result (8.9%) on the wavy shaped electrodes is approaching the value for the PTAA case (5.2%).
| Number | Date | Country | Kind |
|---|---|---|---|
| 1106772.5 | Apr 2011 | GB | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/GB2012/050869 | 4/20/2012 | WO | 00 | 11/21/2013 |
