Characterizing GaN Based Materials Using Modern X-Ray Methods

Characterization of Gallium Nitride (GaN) based materials is often done using non-destructive X-ray metrology. A major class of wide band gap semiconductors, GaN and its In/Al equivalent are utilized on substrates of many new devices like high electron mobility transistors (HEMTs), light emitting diodes (LEDs), laser diodes, and solar in the form of thin film epitaxial layers that range from a few nm to microns. Device performance is frequently linked to stacking sequence, layer thickness, crystal structure, and chemical composition. These properties can be studied using the D8 DISCOVER multi-purpose diffractometer (Figure 1) with techniques like X-ray reflectivity (XRR), reciprocal space mapping (RSM), and high-resolution rocking curves (RC).

D8 DISCOVER experimental setup

Figure 1. D8 DISCOVER experimental setup

Experimental Setup

On the Primary Side

  • Cu X-ray tube of 2.2 kW
  • 2-position primary optical bench
  • Göbel Mirror
  • 004 2 bounce Ge Monochromator
  • Track fix Rotary Absorber equipped with collimator mount

Sample Stage

  • Centric Eulerian Cradle (chi, phi, x, y and z)
  • Vacuum chuck

Secondary Side

  • Universal detector mount featuring a motorized slit
  • PATHFINDER with Scintillation Counter or LYNXEYE XE 0D/1D/2D-mode detector

This versatile system has certain features, like alignment-free component mounting, SNAP-LOCK, and the PATHFINDER equipped with motorized switching between three secondary beam paths. It is able to adjust configurations rapidly and easily to complete the following measurements for complete characterization of GaN based systems:

  • X-Ray Reflectivity (XRR)
  • High Resolution Rocking Curves (RC)
  • Two Theta-Omega scans (2q/w)
  • Reciprocal space mapping (RSM)

Measurement Techniques and Results

X-ray Reflectometry (XRR)

The interference of the specular reflection of X-rays off interfaces or smooth surfaces with different electron densities can be investigated using the X-ray reflectometry (XRR) method. The characterization of roughness, thickness, and density of multi layers, single layers, super lattices etc., with thicknesses spanning from a monolayer to a micron can be analyzed using the XRR method.

In the first example, a thin layer of GaN placed on sapphire by chemical vapor decomposition (CVD) (Figure 2) was used to collect XRR data. The data was collected within 5 minutes. DIFFRAC LEPTOS can be utilized to model and fit the results. It employs Parratt’s formalism to determine reflectivity from these multilayered samples. From the results, it is understood that the deposited film is made up of three layers: a top layer of GaN with 0.92 nm thickness and a lower density, an interfacial layer measuring a thickness of 1.19 nm and a GaN layer measuring 2.69 nm thick.

Modeling and fit results of a GaN film grown on Sapphire using CVD.

Figure 2. Modeling and fit results of a GaN film grown on Sapphire using CVD.

Many complex structures, such as InGaN/GaN MQW (Multiple Quantum Well), with a super lattice barrier grown on 1 µm GaN layer can be characterized using the XRR method. Figure 3 illustrates the overlay of the raw data, shown as a black curve, and the fit data as a blue curve. Super lattice period for both the MQW and the super lattice barrier is also acquired. It is found that the thickness of the complete stack on top of the 1 µm GaN was 219 nm.

Modeling and fit results of a MQW structure with a super lattice barrier.

Figure 3. Modeling and fit results of a MQW structure with a super lattice barrier.

Rocking Curves

Characterization of the structural quality of starting template GaN layer is imperative, as GaN grows in the form of columnar hexagonal crystals that tilt slightly out of plane and have in plane twists on commercially available substrates. GaN serves as a preliminary template for all succeeding GaN layers. Investigations undertaken, using transmission electron microscopy, demonstrate that the in plane twists and out of plane tilts can be respectively related to threading edge and threading screw dislocation densities.

RC can be performed on the GaN layer to calculate the dislocation densities. This method is fairly simple,  an open detector is attached at the predicted 2θ position for a specific hkl, and as the sample is rocked from the negative to positive θ, the diffracted intensity is documented. It is possible to acquire the threading screw dislocation density from the full width at half maximum (FWHM) of a rocking curve carried out on the out of plane reflection of GaN, and the threading edge dislocation density from the in-plane component of an asymmetric plane.

Large deviations were seen in FWHM when the rocking curve method was used to analyze different XY locations of a 4 inch-thick GaN film, produced on silicon (Figure 4 and 5). The observations illustrate that the (004) GaN FWHM differed from 0.073 to 0.092º, and for (102) GaN, the FWHM differed from 0.08 to 0.1º. Based on this analysis, it is evident that the GaN layer does not have a uniform structural quality over the wafer.

(004) GaN FWHM variation across a 4 inch wafer.

Figure 4. (004) GaN FWHM variation across a 4 inch wafer.

(102) GaN FWHM variation across a 4 inch wafer.

Figure 5. (102) GaN FWHM variation across a 4 inch wafer.

Two Theta Omega Scans

Probing coherent scattering of X-rays that originate from interatomic layers is performed, using on axis two theta-omega scans. The interatomic layers are parallel to the surface of the sample, and enable the determination of out of plane lattice constants, Al/In concentration, or check for additional peaks representing unwanted secondary phases. To demonstrate this, an Al(x)Ga(1-x)N film developed on 1 µm GaN buffer layer was examined. An analyzer crystal in the PATHFINDER was used to collect the two theta omega scan. The analyzer crystal is employed to clearly indicate the detector’s angular acceptance, and to decrease the contribution induced by the mosaic spread. Degradation of resolution and loss of vital data, such as thickness fringes from following In/AlGaN layers developed on upper part of the GaN layer are the effects of mosaic contribution. The thickness of the film can be determined using the (002) GaN and Al(x)Ga(1-x)N reflections that are seen together with interference fringes (Figure 6).

(002) Al(x)Ga(1-x)N /GaN Two Theta Omega scan.

Figure 6. (002) Al(x)Ga(1-x)N /GaN Two Theta Omega scan.

Factors like film thickness, chemical concentration, and lattice parameters for an initial sample can be realized using DIFFRAC.LEPTOS, until the raw data and simulation match each other. The layer thickness of the deposited Al(x)Ga(1-x)N is 31.06 nm and its Al content is 23.22% (Figure 7).

Modeling and fit results from (002) Al(x)Ga(1-x)N grown on 1µm GaN.

Modeling and fit results from (002) Al(x)Ga(1-x)N grown on 1µm GaN.

Modeling and fit results from (002) Al(x)Ga(1-x)N grown on 1µm GaN.

Figure 7. Modeling and fit results from (002) Al(x)Ga(1-x)N grown on 1µm GaN.

For this example a model is used that considers the layer of Al(x)Ga(1-x)N to be fully strained, so the position of the peak layer affected alone by the chemical composition. Relaxing or partial relaxing of the Al(x)Ga(1-x)N layer will make the lattice parameters of the layer entirely different. The degree of relaxation, combined with the compositional change measures the layer peak position in the two theta omega scan. The characterization of strain and relaxation in the Al (x)Ga(1-x)N layer using reciprocal space mapping method is essential.

The thickness and concentration data of complex lattice structures, like InGaN/GaN MQW, can be determined using the on axis two theta omega scans. Figure 8 shows the modeling and fit results, where the superimposition of raw data is represented by a black curve and fit data by a blue curve. The data related to thickness is similar to the data acquired previously with the XRR. The analysis reveals that the super lattice barrier, and the average In content of MQW are found to be 6.62% and 14.8%, respectively.

Modeling and fit results from MWQ with a super lattice barrier

Figure 8. Modeling and fit results from MWQ with a super lattice barrier

Reciprocal Space Mapping

Reciprocal space mapping (RCA) constructs a 2D intensity map in reciprocal space, by integrating a set of real space measurements around a Bragg reflection of the film and substrate. The composition and degree of reflection can be measured depending on the position of the reciprocal lattice points of an asymmetric reflection. The commonly used reflection for GaN is 114. In order to ensure that the film is fully strained, the asymmetric reciprocal space point meant for the film should be allowed to lie on a vertical line joining to the substrate. To fully relax the film, the same should lie along the dotted lines joining the source of the reciprocal space, as well as the substrate point (Figure 9).

Visual representation of strain/relaxation in Reciprocal space for asymmetric reflections

Figure 9. Visual representation of strain/relaxation in Reciprocal space for asymmetric reflections

Figure 10 displays completely strained Al(x)Ga(1-x)N grown on GaN layer using the 114 reflection reciprocal space map. Normally reciprocal space maps have been generated with a 0D detector and analyzer crystal aiming for high resolution in reciprocal space. The entire process can take about 6 - 12 hours.

114+ RSM of strained Al(x)Ga(1-x)N on GaN

Figure 10. 114+ RSM of strained Al(x)Ga(1-x)N on GaN

An 1D detector, like the LYNXEYE XE could be employed to decrease the measuring time. A large reciprocal space area can be mapped rapidly by integrating a sequence of 1D fixed scans at varied increments in Two Theta, Omega scans. Figure 11 shows a comparative representation of 0D and 1D (114) reciprocal space maps.

A proprietary film grown on GaAs. 0D RSM - 9 hours measurement time (top) and 1D RSM - 1 hour measurement time (bottom) give the same results.

Figure 11. A proprietary film grown on GaAs. 0D RSM - 9 hours measurement time (top) and 1D RSM - 1 hour measurement time (bottom) give the same results.

Conclusion

This article shows how various X-ray scattering techniques like XRR, RSM, rocking curves and two theta omega scans can be used to characterize single GaN layers to more complicated MQW with a super lattice barrier. Measurement of these layers is easily carried out using a D8 DISCOVER diffractometer, thanks to advanced technologies like the PATHFINDER optic and SNAP-LOCK. The data obtained from these innovative methods is analyzed with the aid of DIFFRAC.LEPTOS, which employs fast and robust algorithms and gives the required quantitative outcomes to define the GaN-based materials.

This information has been sourced, reviewed and adapted from materials provided by Bruker AXS Inc.

For more information on this source, please visit Bruker AXS Inc.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Bruker AXS Inc.. (2019, March 04). Characterizing GaN Based Materials Using Modern X-Ray Methods. AZoM. Retrieved on July 17, 2019 from https://www.azom.com/article.aspx?ArticleID=12658.

  • MLA

    Bruker AXS Inc.. "Characterizing GaN Based Materials Using Modern X-Ray Methods". AZoM. 17 July 2019. <https://www.azom.com/article.aspx?ArticleID=12658>.

  • Chicago

    Bruker AXS Inc.. "Characterizing GaN Based Materials Using Modern X-Ray Methods". AZoM. https://www.azom.com/article.aspx?ArticleID=12658. (accessed July 17, 2019).

  • Harvard

    Bruker AXS Inc.. 2019. Characterizing GaN Based Materials Using Modern X-Ray Methods. AZoM, viewed 17 July 2019, https://www.azom.com/article.aspx?ArticleID=12658.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit