News Article

Possible Low-k Solution And Other Potential Applications

Integration of low-k into a reliable semiconductor production process remains challenging and clearly new ideas are necessary. Seref Kalem of Turkey's National Institute of Electronics and Cryptology describes a cryptocrystalline material with possible low-k and other applications.
There is an increasing demand for low-k dielectrics as inter-level metal insulator for advanced silicon circuit manufacturing. For the 65nm era and beyond that is due in 2007, dielectric values of lower than 3.0 were predicted by 2003 International Technology Roadmap for Semiconductors (ITRS). But, advanced semiconductor manufacturing has additional requirements for low-k such as survival and integration into the whole fabrication process. Cryptocrystalline ammonium silicon fluoride (NH4)2SiF6 (also known as ammonium fluorosilicate) is a low-k material (k=1.87) that we have recently developed on Si-based substrates offers interesting low-cost and high performance solutions to the low-k challenge and for other manufacturing applications. But, there are a number of problems to overcome in getting this material working for the semiconductor industry.

Cryptocrystalline thin films of ammonium silicon fluoride (NH4)2SiF6 (ASF) form on Si when a wafer surface is exposed to a vapours mixture of hydrofluoric acid (HF) and nitric acid (HNO3). The HF:HNO3 mixture is also known as an etching and cleaning solution in semiconductor production. The chemicals used in the process are semiconductor grade 40% HF and 65% HNO3 by weight. The samples in our experiments were made at room temperature using boron-doped p-type (100) Si wafers with resistivities of 5-10Ωcm. But, the ASF crystals were also successfully grown on n-type Si wafers with {100} and {111} crystalline orientations and other Si based substrates such as silicon (Si3N4) and Si1-x Gex( x ≤ 0.3 ).

Vapour phase etch

Growing low-k dielectric layers by vapour phase etching has several advantages:

* no electrical contacts are needed

* selective processing is possible

* layers are homogeneous

* thickness can be controlled

* offers a cost-effective solution to dielectric layer growth compared to other techniques, such as molecular beam epitaxy, chemical vapour deposition and even spin-on technology

Before proceeding further, let us define the cryptocrystalline state. The crystal texture for some materials is so finely grained that no distinct particles are discerned under imaging in a polarisation microscope, and even under scanning electron microscope (SEM), as shown in Figures 1 and 2. The state of matter arranged in this way with such minute crystals is said to be cryptocrystalline or cryptogranular. This type of crystal can exhibit extraordinary dielectric properties that can be used in various fields ranging from microelectronics and packaging applications to photonics and optics.

Fig.1: SEM cross-sectional image (at x3000 magnification) of ammonium silicon fluoride (NH 4) 2SiF 6 cryptocrystals grown on Si from a HF:HNO 3 vapour mixture

The synthesis of cryptocrystals of ammonium silicon fluoride (NH4)2SiF6 on silicon may be an important step forward for the integration of a variety of materials into advanced silicon circuits. With their associated dielectric values, they offer a potential low-k solution. Such layers can also be used as a buffer, bridging the gap between the large lattice mismatches between Si and other materials. With fluoride buffer layers, advanced semiconductor hetero-structures can be directly grown on Si wafers. Thus, inexpensive electronic and photonic devices can be produced and integrated.

An added benefit resulting from the water solubility of fluorides is that the films grown on this buffer layer can be readily lifted-off for some practical applications, such as heat sinking and regrowth. The heat sinking performed in this way may allow laser operation at high temperature.

The vapour phase etching method used in this work can be used for direct formation of advanced lithographic structures and rewritable high-density information storage cells on silicon. The resulting layers from this process can be used as layers for etch stops, device isolation or metallisation diffusion barriers. Moreover, the process itself can provide further insight into the porous structure formation mechanism in silicon, since the same chemicals are involved in material processing.


Figure 1 shows a SEM cross-sectional image of an ASF cryptocrystal layer grown by vapour phase etching on Si. This figure shows how a Si wafer surface turns into an ASF cryptocrystals. The resulting layer is granular, porous and there is no indication of its crystalline nature. The thickness of the layer is around 21µm. The interface between the layer and the wafer surface is relatively smooth. Moreover, there is good adhesion of the layer to silicon. With a closer focus (increased magnification) on the interface, one can see better the quality of the ASF layer which is derived from the Si surface. We should note that the crystal sizes are smallest at the interface. They grow larger toward the surface and become the largest in size at the surface. Figure 2 shows an SEM micrograph of the Si/layer interface.

Fig.2: SEM cross-sectional view of the interface between Si wafer and ASF cryptocrystal layer at x7500 magnification

The layers of ammonium silicon fluoride are probably formed by silicon mediated coupling reactions between HF and HNO3 species on the wafer surface. Unfortunately, a detailed description of the chemical reaction leading to synthesis of ammonium silicon fluoride is not yet known. The reason is mainly due to the lack of evidence for a complete account of the reaction products. In our opinion, not only ammonium silicon fluoride, but also oxygen (O2) comes out as a reaction product. The chemical reaction producing cryptocrystals is probably governed by the following overall reaction equation:

6HF + 2HNO3 + Si à (NH4)2SiF6 + 3O2

Nevertheless, this assumption has yet to be proved experimentally.

Crystal structure, vibrational and surface properties of the layers have been determined through a combination of x-ray diffraction (XRD) analysis, Fourier transform infrared (FTIR) spectroscopic measurements and scanning electron microscopy (SEM) techniques. The annealing experiments were carried out between 50-175ºC and enabled us to study structural integrity and composition of the material.

X-ray diffraction as shown in Figure 3 exhibits sharp peaks indicating that the film has grown in a crystalline structure with a preferred orientation of {111}. The figure shows the x-ray diffraction intensity as counts per second taken from our free standing samples at room temperature. The strongest three peaks are located at 2q=18.34º, 43.14º and 37.14º, where q is the diffraction angle. The corresponding inter-layer spacing-d is estimated using Bragg's law (nq=2dsinl, where l=1.54Å and n=1) to be 4.834Å, 2.096Å and 2.419Å, respectively. The results of these analysis are in agreement with the data reported on cryptohalite crystals of ammonium silicon fluoride (NH4)2SiF6 with a lattice constant of 8.395Å. These results suggest a crystalline character of the layer with the cryptocrystals of ammonium silicon fluoride (NH4)2SiF6 exhibiting isometric hexoctahedral symmetry.

Fig.3: X-ray diffraction pattern of free-standing cryptocrystal layers of ammonium silicon fluoride (NH 4) 2SiF 6 grown by vapour phase etching in a vapour of HF and HNO 3 . Sharp peaks arise from the randomly oriented fluoride cryptocrystals in the film

The as-grown layers have a crystalline white granular, porous structure with a smooth surface. In order to identify the nature of the layer, thermal annealing experiments were carried out on a hot plate. During this process, the surface is exposed to air but not protected against possible decomposition and outgassing. We observed that surface integrity remained intact during thermal anneal up to about 150ºC. However, decomposition of the surface, with a high density of pits forming, starts to occur at temperatures greater than this. Also, at higher temperatures, we observed the formation of individual large fluoride crystals scattered throughout the surface of the wafer. In this particular case, crystal sizes of up to about 100µm can be obtained on the samples. Also, the formation of deep micropores was observed on the Si surface. Note that some parts of the surface were not decomposed, exhibiting a solid compound like structure. This property obviously needs to be investigated further in detail.

Figure 4 shows the typical room temperature FTIR transmission spectrum of an as-grown fluoride layer with wave numbers between between 400cm-1 and 4000cm-1. The presence of interference fringes in the FTIR spectra also indicates a homogeneous film with reasonable optical quality. From the interference fringes in the IR spectrum, we estimated a thickness of 8.1µm using a refractive index of n=1.369 for the ammonium silicon fluoride. (Note that the SEM sample was from a different run and has a thickness of 21µm.)

Fig.4: Room temperature FTIR transmission spectrum of ammonium silicon fluoride (NH 4) 2SiF 6 cryptocrystalline layer on Si. Vibrational bands are of N-H and Si-F stretching modes, typical of (NH 4) 2SiF 6 species

The spectrum exhibits a large number of strong IR absorption bands, in addition to the usual but relatively weak Si-O stretching mode at 1083cm-1. We observed very strong vibrations at 480cm-1, 1433cm-1, 725cm-1, and an asymmetric band at 3327cm-1. The IR vibrational bands observed in our samples are in a very good agreement with vibrational frequencies of ammonium silicon fluoride (NH4)2SiF6 (99.9999%) as an inorganic material. By analogy to previous studies, we have assigned these bands to various N-H and Si-F related vibrational modes. The Si-O vibrations originate from the oxide structure at the interface between the layer and Si substrate. Note that as-grown layers don't contain any oxide as evidenced from the FTIR spectra. The oxide builds up at the interface when the layer is thermally annealed or exposed to air for an extended period of time. This property suggests that the surface should be protected against oxidation in order to avoid strain related problems at the interface.

Another feature is related to hydrogenation effect. The presence of very weak Si-H bands at 635cm-1 and 2125cm-1 in the spectrum of the annealed samples suggests that some hydrogenation effect also takes place with thermal annealing. In this case, the hydrogen probably originates from the HF molecules released from the thermal decomposition of (NH4)2SiF6, thus leading to a hydrogen terminated surface.

There are also other interesting effects with these cryptocrystal structures. Under certain conditions, they grow as very long wires with diameters ranging from few nanometres up to about 1µm widths. One of these features is shown in Figure 5. This particular wire is about 1100nm wide and 40µm long.

Fig.5: Typical wire of ammonium silicon fluoride ranging from a few nanometres to about 1µm (at x10000 magnification) were synthesised by vapour phase etching on silicon wafers. Background structure represents cryptocrystals of fluoride

Key question

Now one can raise the following key question: can we integrate this technology into semiconductor manufacturing? There seems to be promising solutions but, with tough challenges to overcome. Semiconductor companies are trying to integrate a low-k (about 3.0) material for the 65nm node. The lower-k means lower material density and therefore lower mechanical and thermal strength, and lower immunity to process chemicals including water. Thus, the real challenge is to get the material to withstand the various rigorous processes and maintain its properties after the wafer has been manufactured. We believe that this may be possible with greatly improved layer quality - but we will have to wait for the results of the remaining work to decide to what extend the integration and survival of the cryptocrystals can be accomplished.

At the moment, we are looking for funding to achieve the following goals in the area of layer growth, characterisation and implementation of the results. First of all, there is a need to improve and understand the growth mechanism in order to control the process and synthesise thermally and structurally stable layers surviving the entire manufacturing process.

In characterisation, a detailed investigation of structural, optical, thermal, mechanical and electrical properties is required to fully explore the observed physical and electronic effects. On the applications side, we would like to test the performance of these layers as a low-k dielectric insulator in the frequency range of interest; investigate its use as a sacrificial layer in microelectronics and photonics device fabrication including solar cells; explore the possibility of its use as a rewritable high-density information storage media; and investigate other potential applications in nanotechnology, packaging, sensing and information security.

We are also trying to organise a consortium for the upcoming calls in this field to apply for European Framework 6 (FP6) programme support. For the implementation of this low-k solution, a strong industrial support is needed to proceed in making these cryptocrystalline nanostructures a cost-effective high performance solution for the industry.

Acknowledgement: The author thanks the Materials Science Department of TÜBITAK for performing the scanning electron microscopy.


Seref Kalem, TUBITAK - UEKAE, The Scientific and Technical Research Council of Turkey - National Institute of Electronics and Cryptology, Gebze 41470 Kocaeli, Turkey.

AngelTech Live III: Watch the virtual event ON-DEMAND!

AngelTech Live III was broadcast live on 12 April 2021, 10am BST, rebroadcast on 14 April (10am CTT) and 16 April (10am PST) and features online versions of the market-leading physical events: CS International and PIC International PLUS a brand new Silicon Semiconductor International Track!

The event covered the whole spectrum of key developments affecting the compound semiconductor industry. Over the last few months interest in deep-UV LEDs has rocketed, due to its capability to disinfect and sanitise areas and combat Covid-19. This was considered a roadmap for this device, along with technologies for boosting its output.

With 3 interactive sessions over 1 day AngelTech Live III proved to be a key event across the semiconductor and photonic integrated circuits calendar.

So make sure you sign up today and discover the latest cutting edge developments across the compound semiconductor and integrated photonics value chain.


Leybold Launches Small High Vacuum System For Research And Development
Picosun Strengthens Its Position In The Semiconductor Market
SiTime Enables Up To 25% Faster Wireless Charging
SEMI Applauds European Chips Act
Path To Sensors Interoperability - IEEE SA Sensors Series Live Webinar
EV Power Electronics: Driving Semiconductor Demand In A Chip Shortage
Sivers Semiconductors’ Japanese Customer Takes Step Towards Mass Production
Institute Microelectronics And Photonics Chooses Tektronix For Testing
YES Acquires Kanthal’s Semiconductor Equipment Business
Edwards Officially Opens New Flagship Service Technology Centre In Dublin
Merck Launches Semiconductor Materials Webinar Series
Excellent Batch Process Results With Picosun New Generation Tools
Renault To Create Better Brand Visibility For Qualcomm In EV Market
Linton Crystal Technologies Uses AI To Detect Crystal Structure Loss, Notify Operator
ACM Research Enters Bevel Etch Market To Support Emerging Process Steps
Pfeiffer Vacuum Presents A Cloud-based Solution For Service Management
Edwards Opens Lab At Hillsboro Innovation Center
SkyWater Signs Technology Transfer And License Agreement With Deca
New OmniControl Universal Control Unit For Pumps And Measurement Instruments From Pfeiffer Vacuum
Advantest Introduces Evolutionary V93000 EXA Scale SoC Test System
Tektronix Releases Results Of Its Global Engineer Survey 2021
Xenics Expands Its Bobcat Product Family And Makes SWIR Cameras Affordable.
PragmatIC Semiconductor Re-invents Iconic Processor
PragmatIC Semiconductor Secures $80 Million Funding

Search the news archive

To close this popup you can press escape or click the close icon.
Register - Step 1

You may choose to subscribe to the Silicon Semiconductor Magazine, the Silicon Semiconductor Newsletter, or both. You may also request additional information if required, before submitting your application.

Please subscribe me to:


You chose the industry type of "Other"

Please enter the industry that you work in:
Please enter the industry that you work in:
Live Event