Little known MBE facts: High Purity Material

Faebian Bastiman

I was asked recently what purity of material a user should use in their MBE system. Here the grower was referring to the number of “N’s”. The N of the material is a reference to its purity and is actually the number of nines: Material that is 99% pure and 1% impure is termed 2N. 99.5% is termed 2N5, the “2N5” tells you it is at least half way to being 3N. Of course this does not tell you what the impurities are specifically, however you can assume they are made up of all the things you do not want in your III-V thin film layers (Zn, Cu, Fe, Sn, Hg, Ca, Te, etc) and a few III-V elements that you do not need to worry too much about (you are, after all, growing III-Vs anyway).The purity of material will ultimately determine the quality of your thin films, since any impurities in your cell material will likely incorporate as unintentional dopants in your layers.

If you are growing metals (for example MnAs) you are not too worried about unintentional doping and you may use a maximum of 5N5 material (99.9995%). If you are growing electronic or opto-electronic grade semiconductors (for example GaAs) you will want higher purity. Ultimately the highest you can get commercially is 9N (99.9999999%), however I would not recommend you all rush to the shops and buy such expensive material for general research. 9N is exponentially more expensive than 8N, 8N is exponentially more expensive than 7N. Similarly the ultimate background doping you can achieve with 9N is an order of magnitude lower than with 8N, and the same applies when comparing 8N to 7N. Before I suggest the appropriate material quality, let’s do a thought exercise with GaAs…

GaAs has a lattice constant of 0.565338 nm, and therefore an atomic density of 4.42 x 1022 cm-3. When opto-electronics people (specifically detector people) talk about background doping requirements they say that 1015 cm-3 is already good.  This means they would like the unintentional doping level to be 5×107 times lower than the atomic density. In this case 7N5 would be the appropriate choice for you group III and V material. One could argue that you should use 7N5 for all materials and that replacing a specific material with 8N is a waste of money. However this “all or nothing” philosophy is not really justified since unintentional doping is accumulative, and hence replacing your group IIIs with 8N might reduce your background doping from high 1015 to low 1015. Thus 7N5 to 8N is a good choice for opto-electronic research, and 9N is only for exceptional high mobility cases or world record attempts. Moreover, if you are simply experimenting with a new opto-electronic alloy and not so interested in device quality at this stage 6N5 is acceptable. Note that the source material is only one factor in your ultimate achievable material quality, for more information read my Optimum Quality post.

Finally consider the dopant material. When we dope a semiconductor we typically dope in the 1016 to 1019 cm-3 range depending on the application. At 1019 cm-3 the dopant is around 104 times lower in atomic density than the III or V species. This means the doping fluxes are around 104 times lower that the group III and V fluxes and hence any impurities introduced into your system from the dopant source will also be 104 times lower. Add to this that when you dope the alloy you are “trying” to add impurities and you can say that 6N is already a very good grade for dopant material and 5N may even be suitable.

Advertisements

Little known MBE facts: RHEED oscillations (3)

Faebian Bastiman

I was recently reading a nanowire publication and I was reminded of another means of calibrating the group V flux that I used in my III-Sb days. This is the preferred method for III-Sb epitaxy, however  it is also applicable to general III-V growth. In the following I will use GaAs as an example.

The first step is to establish your Ga growth rate using the method in my RHEED oscillations (1) post. Then convert this into atoms/nm2/s using the method in my flux and growth rate post. Finally, you can then calculate the atomic flux of the Ga cell versus temperature using the method in my Arrhenius plot post.

Then set the As cracker to a value that you wish to determine in atomic flux, for example set it to 25 % open. Set the Ga cell temperature to give you a flux of 0.069 atoms/nm2/s [i.e. 0.1 ML/s growth rate on GaAs(100)]. Open the Ga shutter and record the RHEED oscillations in the usual manner. As long as the As flux is larger than the Ga flux you should obtain a growth rate of 0.1 ML/s ± an error of up to 5% depending on how well the RHEED intensity oscillated. Note the error gets larger the fewer oscillations you obtain. Hopefully you can get at least 10 oscillations.

Next double the Ga cell’s flux to obtain a growth rate to 0.2 ML/s and (importantly) leave the As flux set to the original value. You will need to leave the Ga cell to settle for 10 minutes after changing its temperature.  As long as the As flux is larger than the Ga flux you should be able to obtain a growth rate of 0.2 ML/s ± 5%. The magnitude of the As flux compared to the Ga flux is key to this method. Keep increasing your Ga flux until (eventually) the RHEED oscillations no longer yield the growth rate determined by the Ga cell. When this happens the growth rate is no longer dictated by the Ga flux, it is dictated by the As flux. You can check this by increasing the As flux and repeating the measurement that gave the lower than expected growth rate.

If you plot out all your data points you should obtain a graph like the one shown in Figure 1. You can see that starting at small Ga flux, the growth rate initially increases linearly until eventually it becomes As poor (Ga rich) and the growth rate is limited by the As flux. The growth rate you obtain under these As poor (Ga  rich) conditions indicates the As growth rate in ML/s. 0.5 ML/s in this example. You can then convert the As growth rate into an As flux using my flux and growth rate post once more.

Rheed 3 fig

Little known MBE facts: Flux determination

Faebian Bastiman

A flux monitor or monitoring ion gauge (MIG) is a very useful tool for MBE growth.  The principle is rather simple: insert an ion gauge head into the beam path and instead of monitoring the background pressure you can monitor the beam equivalent pressure (BEP). Each elemental species responds differently to ionisation and detection, and the angle of incidence and temperature difference also affect the final measured values. Handily any given cell is in a fixed position relative to the MIG. The ionisation potential is a constant and the effect of temperature differential is small over a 100°C range.

Gathering flux data is straight forward. For an effusion cell (e.g. Ga, In, Al or Bi) you simply need to vary the cell temperature over the range of interest and note down the flux when the shutter is open and when the shutter is closed. Typically the measurement may take 10-60s to stabilise. When you subtract the shutter closed “background” from the shutter open “measured” flux you end up with an “actual” flux for that cell temperature. The As and P cracker or Sb valved sources are quicker. With the source set to operating temperature you just need to step the valve position through the available range. Stabilisation times will vary and to some extent the background flux is moot since the background is predominantly the group V species being monitored.

The result so far is a system dependent BEP determination for each cell. This is useful for you, but not for anyone reading your publication and attempting to repeat the results. What you want is a system independent atomic flux determination. In order to achieve that you need to do a little more work.

The group III cells’ flux can be calibrated in terms of ML/s from RHEED oscillations using Little Known MBE facts: RHEED oscillations (1). This can be converted into atoms/nm2/s using the method in Little Known MBE facts: Growth rate and Flux. We now have a system independent flux in atoms/nm2/s for all our group III species, but what about group V?

Let’s take As as an example. Using the static map you created in Little known MBE facts: Making a static reconstruction map you can estimate the As flux that is necessary to retain a As-rich (2×4) reconstruction at a growth temperature of 580°C. In Little known MBE facts: Group V overpressure we called this flux AsAs. When growing GaAs we also need to balance the incident Ga flux with an additional As flux: AsGa. Thus the total As, Astotal = AsAs + AsGa. This enables a good number of RHEED oscillations (>30) and maintains a (2×4) during growth with very dim 2nd order rods. What happens if you reduce the As flux when growing? Well when you reduce the As to less than Astotal and open the Ga shutter for a few seconds the RHEED will be a slightly weaker (2×4). Close the Ga. Reduce the As flux further. Open Ga for 5s, check the RHEED and close the Ga. Eventually the RHEED will be (1×1), reducing the As further will result in a Ga rich (4×2) pattern (for more information see my (2×4)/(4×2) post). At the point where the (1×1) emerges the AsGa part of the Astotal flux and Ga flux are equal. If you now subtract AsAs that you estimate at this temperature from the Astotal that gave you a (1×1) you have a value of BEP that represents the atoms/nm2/s of As atoms that is exactly equal to the atoms/nm2/s of Ga. So if this growth rate is 1ML/s, that is 6.258 atoms/nm2/s of Ga and therefore also 6.258 atoms/nm2/s of As.

The atoms/nm2/s for P and Sb can be can be calculated using the method outlined for As above, providing of course you have III-P and III-Sb substrates to grow the binary on. The other method is to grow a dilute ternary i.e. GaAsP or GaAsSb and with knowledge of the composition and As fluxes you can estimate the P and Sb fluxes. The problem is you are estimating the flux in atoms/nm2/s “incorporated” as oppose to atoms/nm2/s “incident”. The lattice site competition and other complications that arise when growing group V ternaries mean the incident and incorporated are significantly different. Another means of checking is to “grow” a layer of pure group V, like Sb or Bi, at 0°C and assume the sticking coefficient is unity. The thickness of this layer can be measured with SEM and then you can estimate the incident flux utilising the atomic density of the amorphous film. Though clearly the pyrophoric nature of P precludes it from this kind of experiment.

Little known MBE facts: Making a static reconstruction map

Faebian Bastiman

A static reconstruction map is one of the most useful items in your MBE arsenal. The map charts RHEED reconstructions against group V flux and applied power (or inferred temperature). Most reconstructions are stable over a wide set of temperatures and fluxes, however luckily for us they undergo very abrupt transitions. The map can enable you to return to any flux-temperature reference point with high accuracy even in the absence of any knowledge of the actual absolute temperature. However, with suitable inferences the absolute temperature can be stated to within ±10 °C in most cases.

The actual map is different for every substrate and depends on the number of reconstructions. Undoped GaAs is a good substrate to start with since it has 4 static reconstructions in the flux-temperature range of interest to an MBE grower: i.e. 300 – 650°C. Note that doped GaAs absorbs thermal energy more readily than undoped GaAs, which means for a given applied power doped GaAs is always hotter than undoped. Thus the map must be repeated for the doped and undoped version. That is true for all substrates, though is less obvious for those with narrow band gaps.

The map itself is generated in a several stages. It begins with defining your parameters. For GaAs this range of As fluxes (0 – 300 mil of movement in the As cracker’s needle valve in my case) and the range of your heater power (0 – 100W in my case). The second stage involves gathering flux data. A monitoring ion gauge (MIG) can be used to record the beam equivalent pressure (BEP) for each As needle valve position (see table for example data and ranges). The BEP needs converting into system independent ML/sGaAs or preferably atoms/nm2/s utilising the method in Little known MBE facts: flux determination.

11

The Omicron MBE-STM system has no reliable thermocouple so the heater power is used in place of temperature in the first instance. Most MBE systems possess a thermocouple located behind the heater element that accurately tracks the temperature at a fixed position from the heater element. However both form a suitable reference and either can be used.

The third stage is sample preparation. Remove the oxide as described in Little known MBE facts: oxide removal and deposit enough material (50 nm is sufficient) to create a clear (2×4) reconstruction. This can then be annealed under a lower As flux (0.5 ML/s at normal growth temperature is sufficient) to achieve a flat surface. Cool the sample until a strong c(4×4) is observed and then (with the As flux at 1ML/s) switch off the heater power and periodically check the RHEED until an amorphous pattern is observed and all 1x spots have vanished. You have just deposited an As cap. Once the RHEED is amorphous the As valve can be fully closed. A nominal heating power can be supplied to prevent the manipulator freezing in the presence of the LN2 cooling 0.2W in my case. This is not essential but certainly preferable.  Just make sure the nominal heating does not desorb the cap! Once all As has been purged from the system (typically several hours) the next stage can begin.

The fourth stage is data gathering. Create a 2D table with As flux on the y axis and heater power on the x axis. With no As flux, start heating the sample in small steps (0.05A in my case). Watch for the As cap desorbing with the RHEED and a c(4×4) RHEED pattern emerging. Mark it down in the table. This represents 300±10°C.  Continuing with no As flux, continue incrementing the heater power/current. C(4×4) is stable with no As flux until 400±10°C, at which point on the [-110] azimuth the 2x near the top will transform into a 4x pattern. This is caused because the patches of (2×4) coexist with the c(4×4) on the surface. The is the c(4×4)/(2×4) transition, mark it as “mix” for short. Continue with no As flux until (2×4) replaces the c(4×4). The (2×4) will remain until around 475±25°C where a (nx6) pattern emerges. The pattern is weak and it is unclear whether n is 3 or 4. At this point, to avoid damaging the sample, open the As to the first position (10 mil in my case). The surface should immediately become (2×4) once more.

At this point keep the heater power fixed and increase the As flux in small steps. The (2×4) will mix once more at an As flux of ~0.8ML/s and will thereafter give way to c(4×4). Continue to chart out the range of fluxes and temperature and eventually the data should resemble the table below.

The fifth and final stage involves interpretation of the raw data. We have already marked 300, 400 and 500°C with some error bar. The earlier oxide remove power can be marked as 600±20°C. Plotting the power vs temperature allows the other temperatures to be extracted from a line of best fit. The temperatures from ~400 to ~540°C can now be readily located with a known As flux by utilising the c(4×4)/(2×4) “mix” transition. If your aim is publication or dissemination you can now create a graphical plot of your data, though for personal use the table retains greater fidelity.

Little known MBE facts: In growth rate

Faebian Bastiman

InGaAs QW and InAs QDs are popular active regions in opto-electronic devices. In Little Known MBE facts: RHEED oscillations (2) a concept was introduced to establish the In growth rate from a known Ga growth rate but what about an accurate and independent determination of the In growth rate? Well, conveniently there are several options.

The first method is simple. All you need is a rather expensive InAs(100) substrate and its lattice parameter: 6.05840 Å. RHEED oscillations can be performed for In on InAs in an identical manner as for Ga on GaAs. The required As flux for maximum RHEED oscillations for any given ML/s growth rate should be ~87% the As flux needed for GaAs; this maintains the 1.6:1 ratio on InAs. If you don’t know why read Little known MBE facts: Crystal hopping. One thing with InAs is that the demand for As over pressure rises significantly with increased substrate temperature. So if you need more than ~1.6:1 it implies that the substrate temperature is too high.

The second method uses the idea of adding and subtracting RHEED oscillation growth rates for ternaries introduced in Litte Known MBE facts: RHEED oscillations (2). In this way the growth rate of GaAs and InGaAs can be used to find the InAs growth rate. To test, try growing a QW of In0.06Ga0.94As of 10nm total thickness and employ a little maths and logic: In0.06Ga0.94As is effectively 9.4 nm of GaAs and 0.6 nm of InAs deposited at the same time. If you use a GaAs growth rate of 0.94 nm/s (for example), the whole QW should take 10 seconds to grow. An InAs growth rate of 0.6 nm/s will then give you the desired thickness and composition. If not, and you are certain about your GaAs growth rate, the InAs growth rate is wrong. As an added check of composition-thickness, a simple 3QW structure can be grown as shown in Figure 1 (below). The RT PL peak wavelength varies as a function of both InGaAs thickness and composition: e.g. [In] = 6%, 10nm QW should give RT PL ~900nm.

sdf

The third method is very useful for very low growth rates. It involves using the S-K transition which occurs when ~1.8ML of InAs is deposited on GaAs(100). There is an abrupt RHEED transition from streaks to spots which is highlighted in figure 2 (below). The actual S-K transition is a little imprecise since it relies heavily on temperature (particularly where non-unity In sticking coefficients are in question). It also is less accurate at higher growth rates, since the exact moment of the transition is subjective and at >0.25ML/s the timing depends on how accurate you are with your stopwatch. You will certainly want to make 3 attempts and average the times. Luckily 1.8ML of InAs can be readily flushed from GaAs with a brief (5-10 minutes) anneal at ~600°C under As flux and then you can cool the sample back to ~500°C and try again.

The fourth option is a clever trick you can play if you have access to a phosphorus source. The III-V alloy In1-xGaxP is lattice matched to GaAs at room temperature for x = ~50%. This little fact means you can grow 250nm of In0.5Ga0.5P/GaAs(100) and then analyse the exact composition with XRD peak splitting. When the Ga and In growth rates are identical the composition will be exactly In0.5Ga0.5P (just make sure In sticking is unity by growing around 480°C). Also if you are using RHEED oscillations you can accurately determine your Ga growth rate with GaAs RHEED oscillations, then work out your InGaP growth rate from RHEED oscillations AND discover the optimum III:V ratio for phosphorus growth at the same time. The fact that one method utilises 250nm of material and the other utilises 30ML means you can also get an idea of how those pesky shutter transients are affecting things.

The fifth and final option is a little super-lattice (SL) on an InAs(100) or a GaSb(100) substrate. Ideally, you will need access to antimony (Sb) and that way you can grow an GaSbAs/InAs(100) or a InAsSb/GaSb(100) lattice matched SL. Lattice matching at room temperature means the substrate peak and SL-average-composition peak coincide and will only see the satellite peaks. You can also utilise binary super-lattice InAs/GaSb but you may get strange effects from the interfaces unless you know what you are doing (more on that later). This technique relies on prior knowledge of the Ga growth rate (again) but can also provide interfacial and structural quality information.

The actual method(s) you use is largely a matter of personal preference and intended application. If you are trying to work out the growth rates so you can grow InGaP/GaAs(100) lattice matched bulk it makes little sense to utilise a InAs/GaSb SL to work out the growth rates. The nice thing is that with so many ways to double-check the actual growth rate should (in theory) be very accurate indeed.

Little known MBE facts: Crystal hopping: GaAs and InP

Faebian Bastiman

GaAs and InP are two very popular material systems for opto-electronic devices. After measuring and calculating and cross-referencing your Ga, In and Al growth rates on GaAs(100) with a variety of techniques the time has finally come to grow on InP(100). Hurray. However for some reason the growth rates have gone awry! The ML/s growth rate you get on InP is 108% what your very careful and accurate measurements tell you is 1ML/s on GaAs. Worse your µm/h growth rate is 112% out. How can I quote your growth rate error with such startling accuracy?

Well, the fact is there is no error in the growth rate. The discrepancy comes because GaAs and InP have a lattice constant of 5.65338 Å and 5.86860 Å, respectively. When you are talking about 1ML/s on GaAs you are talking about a flux of 6.258 x 1014 atoms/cm2/s. In fact you should say 1 monolayer of material with the lattice parameter of GaAs per second or for short: 1 ML/sGaAs. Since 1ML/sInP is in fact 5.807 x 1014 atoms/cm2/s, when you put down 1MLGaAs you are in fact depositing 1.08 MLInP. The ML growth rate of different material systems is proportional to the square of the lattice constant: 

The square term comes from the fact you are only worrying about a 2D ML on the (100) plane, and hence you are only concerned with the ratio of area of the face of the zinc blende unit cell. The larger discrepancy on the µm/h growth rate comes from the fact you are now looking at a ratio of the volume of the zinc blende unit cell. Thus the term is proportional to the cube of the lattice constant: 

This is a good example why the growth rate should be quoted in terms of the substrate independent unit of atoms/cm2/s.

Little known MBE facts: Growth rate (2): Super-lattices

Faebian Bastiman

Superlattices (SLs) were first discovered by their rather unique X-ray diffraction properties, and so it should come as no surprise that SL creation and XRD characterisation continue to go hand in hand. Rather than get into a semantic debate about where a SL starts and a multi-quantum-well MQW structure ends let’s concentrate on designing a useful structure to determine our growth rate. We want the XRD plot to have well defined satellite peaks so we will want >15 periods and we probably want to average over at least 4 satellite peaks (2 negative and 2 positive) so we will want a period of >10nm. A useful test structure is shown in figure 1 (below) which has the SL nomenclature [GaAs10AlAs10]15.

The 004 ω-2θ scan of this structure is shown in figure 2. A simple peak splitting and periodicity analysis can tell us our growth rates. So what we have just done with 300nm of deposited material is to calibrate both the Ga and Al growth rate. You can do that with 30ML of material and RHEED oscillations. So why bother?

 Well the fact is the SL is telling you more. Firstly, you can vary the thickness of the two layers, whilst even maintaining the same period if you wish, and that will allow you to establish exactly what the shutter transients are doing. Secondly the quality of the XRD can give indications on the interfacial roughness and structural quality of the SL. Thirdly, you can also do RT-PL on the structure which can quickly tell you the width of the GaAs layer from the peak wavelength (just ensure the AlAs is thick enough to avoid wave-function over lap!) and that can allow you to gauge the opto-electronic quality of the GaAs:  a useful little test structure indeed.