Density of state or band structure energy can be obtained using any theoretical approach, for example DFT. So, from these results, is it possible to determine the type (n or p) of a given semiconductor?
Dear Cory Cress,
Your answer is generally okay. But you stated in starting your answer that:The DOS contains no information about the carrier density in a semiconductor, and near the end you stated that: to determine you carrier type, you need 1) the DOS and 2) the location of the Fermi level.
There is contradiction between the two statements. The second statement is the correct one. To make this point more clear i would like to add some details.
The density of electrons n in the conduction band of a semiconductor by definition is the integration of the product of the density of states function N(E) and the Fermi function f(E) from the conduction band edge to the top level in the conduction band. The same holds for the holes p in the valence band. To get p, one has to integrate the product of the density of states function of holes in the valence band N(E) by the probability of occupation of the states by holes which is 1-f(E), over the valence band energies starting from the top energy level in the valence band.
If n=p the material will be intrinsic. If n>p the material will be n-type and vice verse.
The position of the fermi level determines the ratio between n and p. When n=p=ni the fermi level coincides with the intrinsic fermi level. If n/p >1 the the fermi level shift upward to the conduction band edge and vice verse for ptype material.
Practically the material is made p type or n type by intentional doping of the suitable impurity as we know .
In principle, yes. However, many important materials have native defects that serve as dopants that can in practice determine the type of the material
For a semiconductor material to be n or p type, experimental procedure will be the best and final option.
You can use current/voltage and capacitance/voltage characterization.
For an n-type material, applying -ve voltage is a reverse bias which will show capacitance drop as the bias voltage increases. On the other hand, applying +ve voltage on P-type material also implies reverse bias which is capable of increasing the depletion thickness as the voltage increases.
If you are using the material for the purpose of detection of radiation, nuclear spectroscopy is another good way of determining the nature of the material.
Existence of the energy state in a particular energy determines that the energy state could be filled by an electron (with possibility dictated by the Fermi distribution function). On the other hand, when at a given energy level there is no state, i.e. DOS=0, it means this state is always empty. so if between two energy level, lets say E1 and E2, DOS is constantly zero, there is an energy gap equals to Eg= E1-E2. if E1-E2. If Eg is zero material behaves like a conductor and if Eg is not zero, depending on the absolute value of this gap, material is semiconductor or insulator in which E1(
Yes, calculated DOS can easily inform about N or P character of semiconductor. The DOS charts show position of fermi level as well as partial DOS from which character of semiconductor can be predicted. If Ef is located near CBM, then it may be N- and if it lies near VBM then it is P-type material.
In my opinion, its much easy to identify the conductivity type by mean of the band structure.
The DOS contains no information about the carrier density in a semiconductor. Lets take Si for example, p-type and n-type silicon have the same DOS but have different majority carrier type. It is the location of the Fermi level, with respect to the DOS, that provides information about a semiconductor's carrier type. At mid-gap, the semiconductor is intrinsic. When its close to the valence (conduction) band it is p-type (n-type). Therefore, to determine you carrier type, you need 1) the DOS and 2) the location of the Fermi level.
In general this information is provided by DFT, but its important to understand the difference between the intrinsic (DOS) and the extrinsic (Fermi level) properties of the material.
Dear Cory Cress,
Your answer is generally okay. But you stated in starting your answer that:The DOS contains no information about the carrier density in a semiconductor, and near the end you stated that: to determine you carrier type, you need 1) the DOS and 2) the location of the Fermi level.
There is contradiction between the two statements. The second statement is the correct one. To make this point more clear i would like to add some details.
The density of electrons n in the conduction band of a semiconductor by definition is the integration of the product of the density of states function N(E) and the Fermi function f(E) from the conduction band edge to the top level in the conduction band. The same holds for the holes p in the valence band. To get p, one has to integrate the product of the density of states function of holes in the valence band N(E) by the probability of occupation of the states by holes which is 1-f(E), over the valence band energies starting from the top energy level in the valence band.
If n=p the material will be intrinsic. If n>p the material will be n-type and vice verse.
The position of the fermi level determines the ratio between n and p. When n=p=ni the fermi level coincides with the intrinsic fermi level. If n/p >1 the the fermi level shift upward to the conduction band edge and vice verse for ptype material.
Practically the material is made p type or n type by intentional doping of the suitable impurity as we know .
Using the theoretical DFT calculation to calculate the DOS including the low level dopants is nearly impossible due to very large number of atoms needed. I don't think DFT calculation under very low concentration of dopants can be done using today's computing capability.
Yes, it does. As others poined out, if you have DOS measured on both sides of E_F, the asymmetry of the band gap around E_F (zero setpoint in many spectroscopies) tells you, what type the material is. On the other hand, different spectroscopies produce different artefacts and you should consult an expert in that specific field with your data, not give in to appearances.
Yes in principle. But the results depend upon the deep states which may pin the Fermi level.
To extend on the answers before, I think it is important to notice that the notion of "p-type semiconductor" (or n-type) is often times used slopply to indicate different behaviors which can be incompatible with each other! This can be observed in particular in the investigation of novel semiconductors (organic, solution processed etc.)
In literature a "p-type semiconductor" can refer to:
1) a semiconductor with a fermi energy closer to the valence band then to the conduction band.
2) a material in an FET that shows p-channel conductivity (i.e. conducts at negative gate voltages). This can even be metal with a low density of states...
3) a semiconductor that forms a solar cell in connection with a "known n-type" material.
Sadly, often times the observation of any of 1,2,3 is used to claim some general "p-type"ness and is implicitly used to infer the other behaviors. This is often times incorrect.
Example:
- A material X has a fermi energy close to the conduction band (therefore n-type by (1)) due to an intrisic defect close to the CB.
- The same material X can show p-type conduction in an FET, if e.g. the same defect strongly reduces the electron mobility while hole mobility stays high.
- The same material X can work as either a n or p-type material in a solarcell depending on the relative bandoffsets to the other material of the hetero junction.
While in classical, well understood semiconductors - such as Si, GaAs, etc. - there is clear relations between these behaviors, they can be very well be decoupled in more complex semiconductors!
Best Regards,
Deniz Bozyigit
I think the best way to measure the amount and type of carrier is by Hall effect. With this technique It is possible to determine both the polarity and type number of carriers.
Great thanks for all intervenant. All the given responses and ideas are very interesting. Thanks again.
I agree with the responses given by Deniz Bozyigit and Daniel Valim.
But there is one easy method to determine the type of semiconductor material.
--> When the two points on the semiconductor wafer are maintained at suitable different temperatures, then an emf is generated across the two points. Just from the polarity of the emf generated we can find out type of the semiconductor material.
If you compute DOS using ab initio technique within the super-cell model, you can observe sharp peaks in DOS which correspond to inpurities' states. If these peaks are closer to the conduction band edge, there is a big chance that you have an n-poped semiconductor. In the case, when the peaks are closer to the valence band edge, you deal with p-doped material.
I think, that the answer is firmly bond to the nature of the semiconductor considered, the doping level and also if it is or not an heterostructure where energy bands are very entangled.
The better way to find the carrier type is Hall Measurement.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Semiconductor