Gallium Arsenide: Another Player in Semiconductor Technology

This article looks at gallium arsenide, and explores how it compares to other popular semiconductor materials, and explores the different components utilizing each material.

Silicon has long held its place as the key material in semiconductors. However, gallium arsenide, along with other compounds like gallium nitride and silicon carbide, are now sharing the stage. So what is gallium arsenide and how does it differ from other compounds? Let's explore this compound and take a look at how it's being used as a semiconductor material.

What is Gallium Arsenide?

Gallium arsenide (GaAs) is a compound built from the elements gallium and arsenic. It is often referred to as a III-V compound because gallium and arsenic are in the III group and V group of the periodic table, respectively.

 

Figure 1. The gallium arsenide compound. Brown represents gallium and purple represents arsenic. Image courtesy of Shandirai Malven Tunhuma - University of Pretoria.

 

The use of gallium arsenide is not a new technology. In fact, DARPA has been funding research into the technology since the 1970s. While silicon-based technology has been “the backbone substance of the microelectronics revolution, GaAs circuitry operates at the higher frequencies and signal amplification powers that have made practical a world connected by palm-sized cell phones.” 

Gallium arsenide led to the miniaturization of GPS receivers in the 1980s. This made the laser-guided, precision munitions that entered US arsenals during that time period possible.

Bandgaps in Different Semiconductor Materials

Without getting into deep theoretical physics, a material’s bandgaps the space between a material’s atomic shell layers. The larger space means that it takes more energy to get the semiconductor's electrons“jump” to the next shell and to make the semiconductor shift into in conductive state. As we shall see, this has a number of important ramifications.

Comparing GaAs, Si, SiC, and GaN Bandgaps

With high electron mobility, semiconductor devices built of GaAs can function at frequencies in the hundreds of GHz.

While not truly considered a “wide bandgap” material, GaAs does have a considerably higher bandgap than silicon does. Critically, this makes GaAs highly resistant to radiation and therefore a great choice for defense and aerospace applications. Another selling point is that GaAs devices are far more resistant to heat and give off less EMI.

GaAs features a direct bandgap as opposed to silicon’s indirect bandgap. Because of this, GaAs can emit light much more effectively than can those made of silicon. This gives GaAs LEDs a clear advantage over those constructed of silicon. 

 

A major advantage of silicon is that in the real world of mass manufacturing, silicon is far easier to work with. Silicon has a “native oxide,” silicon dioxide (SiO2). This ready insulator is an invaluable asset in fabricating silicon devices. GaAs has no analog. 

 

At this writing, silicon processes down to the seven-nanometer level are being developed. 500 nanometers is about as low as GaAs can go at this time. And while GaAs is fast, it takes power. So, for ordinary mid- and low-speed logic, silicon may still be the way to go

Gallium Nitride and Silicon Carbide

As detailed below, silicon carbide (SiC) and gallium nitride (GaN) feature bandgaps that are considerably in excess to those of silicon or GaAs.

 

Comparison of Bandgaps

Material	Bandgap
Silicon (Si)	1.1 electronvolts (eV)
Gallium Arsenide (GaAs)	1.4 electronvolts (eV)
Silicon Carbide (SiC)	3.0 electronvolts (eV)
Gallium Nitride (GaN)	3.4 electronvolts (eV)

 

Silicon carbide can be employed to build power MOSFETs for high voltage, high power applications operating at high frequency. They can tolerate high temperatures and feature RDS (on) values that are stable with temperature. RDS is the resistance from drain to source, an extremely critical parameter in any power application.

 

Figure 2. Silicon carbide. Image (modified) courtesy of the University of Munster.

 

Gallium nitride has an even higher bandgap than silicon carbide and higher electron mobility, too. The technology’s inherently lower output and gate capacitances further enable high-speed operation. GaN devices lack the body diode that is inherent in silicon-based devices. This serves to eliminate recovery loss, increase operational efficiency, and reduce EMI.

 

Figure 3. Gallium nitride. Image courtesy of the University of Bristol.

Silicon carbide can be employed to build power MOSFETs for high voltage, high power applications operating at high frequency. They can tolerate high temperatures and feature RDS (on) values that are stable with temperature. RDS is the resistance from drain to source, an extremely critical parameter in any power application.

Gallium Nitride has an even higher bandgap than silicon carbide and higher electron mobility, too. The technology’s inherently lower output and gate capacitances further enable high-speed operation. GaN devices lack the body diode that is inherent in silicon-based devices. This serves to eliminate recovery loss, increase operational efficiency, and reduce EMI.

The LMG3410R050 GaN Device from Texas Instruments

TI’s approach is to include gate driver circuitry along with a 600V GaN transistor. The LMG341xR050's (PDF) inherent advantages over silicon MOSFETs include ultra-low input and output capacitances for high-speed operation. Switching loss reduction through zero reverse recovery is another benefit.

 

Figure 4. The LMG3410R050. Image courtesy of Texas Instruments.

 

GaN devices like the LMG3410R050 have no reverse recovery losses because, unlike silicon MOSFETS, there is no PN junction between source and drain. 

The integrated gate driver is specially tuned to the GaN device for fast driving without ringing on the gate. It saves time, space and BOM costs for OEM’s and protects against faults by providing over-current and over-temperature protection.

Cree’s Billion Dollar Commitment to SiC MOSFETS

In the world of high bandgap semiconductors, SiC is another powerful contender, as evinced by Cree’s commitment to the technology.

Cree offers many SiC MOSFETs, including the C2M0045170D. This device is rated at 1700V and 72A. Maximum junction temperature is 150°C. Importantly, it sports an RDS (on) of only 45 milliohms.

The company’s CAB450M12XM3 (PDF) is a 1200V, 450A silicon carbide half-bridge module.

 

Figure 5. The CAB450M12XM3. Image courtesy of Cree-Wolfspeed (PDF).

 

Continuous junction operation at 175°C is possible. This high-power device is designed for:

• Motor and traction rives
• Fast chargers for vehicles
• Uninterruptible power supplies

Gallium Arsenide LEDs

These devices are more commonly offered in wafers, but Vishay offers the TSUS4300 (PDF), a discrete GaAs LED radiating at 950 nanometers. One of their specifications is that they offer “good spectral matching with Si photodetectors,” presaging the central point of our next section below.

Is Gallium Arsenide a Better Choice than Silicon?

We’ve discussed some generalities and overall characteristics, but designers have to carefully analyze the particular needs of specific designs and not make their material choice based on preconceived notions. Sometimes, the answer won’t be what was initially expected.

In an article written by Analog Device’s Theresa Corrigan, N-channel CMOS MOSFETs are contrasted with GaAs devices when serving as wideband (900 MHz an higher) electronic switches.

 

The Advantages of GaAs

• Low on resistance
• Low off capacitance
• High linearity at high frequencies

 

The Advantages of CMOS

• Loss of 3dB or less at 4 GHz 
• Low power consumption 
• No requirement for dc blocking capacitors 
• High isolation between ports

 

Her conclusion? Sometimes using the more traditional choice of silicon for devices still comes out on top. However, each new semiconductor material offers different benefits.