I must disagree with some of the comments above. You need to consider what information is desired as well as the nature of the sample and set kV accordingly. Secondary electrons are generated near the surface of the sample and provide topograghic information. High kV results in most primary electrons traveling deeper within the sample, with how deep depending on the sample density and elemental composition. Yield of secondary electrons will be reduced and surface detail will be lost in this case for most samples. On the other hand, back-scattered electrons come from below the surface so a higher kV would be used to increase the yield. Likewise, X-rays come from very deep within the interactive volume and thus high kV is used for EDX even though the image resolution will not be as good.

In the case of a low atomic number samples, as life science samples or polymers, usually it is good to start at 5kV. In fact, now with beam deceleration available on many new instruments, it is not unusual to have landing energies much below 5kV to get both good structure and minimize charge accumulation in the sample. In the case of high atomic number samples such as metals, you may go to 7-10kV for imaging but it is rare indeed that you need to go higher for good secondary electron imaging. The loss of secondary electron signal more than out ways the theoretical prediction that higher kV always results in better resolution.

Try it and see. Start low and then go to much high kV and compare the results. Don't be afraid to try different settings with different samples as this is the key to really good microscopy. Change kV just as you should change probe current (condenser lens strength), working distance, aperture size, etc to adjust conditions to the specific sample with which you are working.

It is basically related to what you are for, Dear Vladimir. If you are interested in finding a feature, basically you have to optimize the parameters like voltage, spot size and magnification. For a general bulk/powder sample( depending on what your sample is), imaging has been mostly done at 15 KV. If you increase your KV, energy of both back scattered electrons and SE will be more. Needless to say, especially you ll get more Z contrast. But if the signal is not sufficient you can still go for 20-25 KV. For a publication it is always better to do imaging under similar conditions.

As Bharathi pointed out, this is really a question of your structure you want to see. Generally a higher voltage results in a better resolution. At the same time your electrons travel deeper into the substrate and you get a signal from a greater depth. I normally use 20-30 kV for ultra-high resolution on bulk samples. For thin nanowires with 10-100nm size I switch to 5 kV to prevent contrasts from underlying nanowires. This results in a small loss of resolution, but in summary the image is clearer. In my opinion it does not need an explanation in a publication, when you just image your sample. Only when you use your electron beam as an active component, like in electron beam induced current, you should explain your choice of voltage.

In general, the choice of acceleration voltage depends on (i) problems you have to resolve; (ii) type of SEM you have; (iii) type of material and preparation technique employed. For example, if you have conventional high-vacuum SEM with tungsten or LaB6 cathode and equipped with EDS system, an acceleration voltage between 15-20 kV is good for EDS analysis, while for high resolution imaging of materials in secondary electrons, acceleration voltage of 25-30 kV is more appropriate.

If you have FESEM (SEM with field emission gun), you have much more possibilities for analysis and imaging, including operations at low voltage 2-5 kV.

I must disagree with some of the comments above. You need to consider what information is desired as well as the nature of the sample and set kV accordingly. Secondary electrons are generated near the surface of the sample and provide topograghic information. High kV results in most primary electrons traveling deeper within the sample, with how deep depending on the sample density and elemental composition. Yield of secondary electrons will be reduced and surface detail will be lost in this case for most samples. On the other hand, back-scattered electrons come from below the surface so a higher kV would be used to increase the yield. Likewise, X-rays come from very deep within the interactive volume and thus high kV is used for EDX even though the image resolution will not be as good.

In the case of a low atomic number samples, as life science samples or polymers, usually it is good to start at 5kV. In fact, now with beam deceleration available on many new instruments, it is not unusual to have landing energies much below 5kV to get both good structure and minimize charge accumulation in the sample. In the case of high atomic number samples such as metals, you may go to 7-10kV for imaging but it is rare indeed that you need to go higher for good secondary electron imaging. The loss of secondary electron signal more than out ways the theoretical prediction that higher kV always results in better resolution.

Try it and see. Start low and then go to much high kV and compare the results. Don't be afraid to try different settings with different samples as this is the key to really good microscopy. Change kV just as you should change probe current (condenser lens strength), working distance, aperture size, etc to adjust conditions to the specific sample with which you are working.

Thanks for all the answers.

I asked my question because I was tired seeing in publications micrographs in secondary electrons (SE) taken at much higher than needed voltages. I was wondering if someone can justify “high voltage way”. I even made my small “research”, investing in it one hour of my time. Using Google Scholar and PubMed I looked through latest published papers with SEM micrographs. I used only available full-lengths papers with only SE imaging (no BSE and EDS); with low and intermediate magnifications (not higher than x10,000); and utilizing only relatively new microscopes (manufactured in current millennium), which are capable of producing nice images at at least 5 kV. Out of 22 such papers only in 6 voltage was 5 kV or lower, and 16 papers utilized voltage 10 kV and higher, mostly 15 and 20 kV (and two of them even 25 and 30! kV) Of course, “high voltage” papers (73% of all papers) had micrographs of worse quality.

Fortunately among RG members (as I see from answers) nobody is utilizing higher voltages just for SE imaging, only in conjunction with other signals (BSE, X-ray). Of course I agree with Debra that lower voltages are a right way to go. I just think she is a bit too generous suggesting 10 kV for metals: for example at 10 kV fine river patterns (brittle fracture) are becoming less visible (more transparent) than at 5 kV or lower. Even highest resolution is not always requires highest voltages, it depends very much on specimen; signal/noise ratio and increased transparency of small features may require to go with lower voltage. Besides resolution there are very few justifications of utilization of higher voltages. One of them (and I use it quite often) is desire to see some of BSE contrast on SE micrographs. If specimen has two (or more) phases with different Z, then at voltages that generate enough backscattered electrons SE micrographs along with topographical contrast will show contrast between phases. Can be useful; the price – some loss in topography.

Generally speaking, the higher the accelerating voltage the greater the resolution. However, higher accelerating voltages frequently result in a phenomenon known as "charging" in which there is not a nice even flow of electrons emitted from the surface of the specimen. It results in a poor and unacceptable image, but if the specimen is adequately coated with a metal, then the issue of charging is minimized. Hope this has been helpful.

Try the following to see the difference. Take a TEM grid with a Formvar layer on, coated with 20 nm of AU/Pd. Start with 1kV and then with 10kV on the SEM. At 1kV you will only see the Formvar film. When the accelerating voltage goes up, the SE electrons are generated at a deeper level of your sample. At 10kV you wouldn't see the Formvar layer but only the TEM grid. So, with biological samples use 5kV or less to get more surface info. With higher kV, the chance of damaging your sample is bigger.

Louwrens ,

Thanks for interesting input. Your suggested a really nice demonstration specimen for students. I think I will use it.

I work with human blood samples and use an inlens FESEM and kV of 1. Our micrographs are particularly clear and for surface work up to 150 000 machine magnification - coating with carbon, otherwise you view the gold particles on the surface. To view some of the micrograph quality just do a pubmed search on E Pretorius + fibrin or red blood cells.

Dear Etheresia,

Unfortunately I have to disagree with your papers. I have seen two of them (one in Microscopy Research and Technique, I think you have referred to that one). Magnification you used was not 150,000, but ten times lower, 15,000. This is too low mag to see gold islands. Anyway, today gold is considered not suitable for high magnifications anyway, and Au-Pd, Pt, Os etc. are mostly used. But at x15,000 any coating will be good enough for observation. Carbon, due to low secondary electrons coefficient, is the least suitable of all coatings you used. Moreover, microscopes with in-lens detector are capable of nice observation of most specimens without any coating at voltages around 1 kV (I suppose you used one of these).

I agree with much of what has already been said. The accelerating voltage will vary widely depending on the type of material (ie: a highly insulating ceramic materials versus a metal) as well as the geometry (specifically, thin film versus bulk versus powder). When I am looking at a new material or a new sample geometry, the accelerating voltage is one of the first things that I play around with in the SEM. I would recommend this to you as well. In terms of publications, use whatever accelerating voltage you want, and just state it somewhere in the paper.

Unfortunately many SEM operators are sticking to the old ways of using high kV; around 15kV is still very commonly seen on images, but this stems from just not being bothered to play around with the accelerating voltage and the days of pre-FEG systems when probe brightness and coherence at low kVs was poor. It is now generally accepted that for life science specimens this is way too high and although one would expect to get higher resolution at higher kV, in fact, due to the greater interaction volume at higher kV, you get lower resolution of the SE signal on biological specimens. This is because you would be confusing the image with more secondaries coming from beneath the surface as well as those from the true surface. So, whether you have beam deceleration available or not, you should use lower kVs on life science specimens to get better resolution. Typically never go higher than 3kV. But it does depend a bit on specimen. For highest resolution I have found about 1kV was optimal for imaging the true surface detail on cryo-specimens of bacteria, using a FEG with an upper detector and with the shortest working distance possible. Higher than than and I start to lose the surface detail and things look slightly transparent and fuzzy due to SEs coming from beneath the surface. It's a different world to materials science specimens.

Then can any one comment regarding the kV, Vacuum inside the chamber etc., to observe thin films of around 50 to 100 nm of semiconductor material on a FTO coated glass substrates.

Thanks in advance.

I used to look at insulating thin films of 100-300nm in thickness on silicon substrates, and I used to get pretty good images without charging at ~8keV. Your materials are a bit different, but you could start out in that area of accelerating voltage and then play around with it. Good luck!

Thank you claire weiss... I will do the measurements and comment on the results.

Sitaramanajaneya, nobody can give you exact parameters for your task. What do you need to see – surface details, cross sections, what is conductivity of your specimen? What are capabilities of SEM you have access to? Etc… Start with the lowest accelerating voltage possible (10 V for Magellan, 2 kV for SEM515), then try higher voltages, and then you can choose one the most suitable for your task. Lower voltage = best surface details (but you need suitable resolution), older microscope = higher voltage if you need highest resolution. A lot of other things to consider like charging, need of BSE or EDS.

Another possibility of getting high resolution images of sensitive specimens is to use a microscope with beam deceleration technique. We just got such a system from Tescan and I tested it the last days. The mechanism behind it is a high voltage bias of -5kV on the stage and thus on the sample. When using 6kV accelerating voltage, one gets a highly focused beam which is decelerated above the sample to 1keV energy and images only the surface layer. I also aquired nice images at 300eV which was not possible before.

ZEISS SUPRA™ 40 Field Emission Scanning Electron Microscope (FESEM) ... This is the instrument I work on and after getting my results by operating under different Voltage conditions will give my conclusions and ask for some help if needed. thank you.

In SEM labs exist at least two types of operators. People like Vladimir or Debby, who knows very well how to operate with microscope. They know background physics and they have a lot of experience. For them this question "what accelerating voltage should be used" is the same like "how I should cut my hairs" ... common simple answer dosn't exist ... So what should do another operators, young PhD and students, people who just started their adventure with electron microscopy after 1-3 days training ?? At the beginning they have to minimize number of parameters. HV, beam diameter, apertures, WD, detector types .. there is so many combinations that they can be lost.

I often suggest to use:

- 30kV for STEM or EDX on bulk metallic samples

- 15kV on older system (W, LaB6, basic detectors) and if You want to work with EDX or BSE - any sample excepct non conductive

- 5 kV on new one (except EDX work) with in lens system and BSE conversion detectors almost on all samples

- 1 kV for non conductive and biological samples

This is just starting point.

For secondary electron images the smallest is the volume of interest the lowest may be the accelerating voltage and the lighter is the element the truer is that sentence. I used often CASINO to simulate trajectories of the beam inside the sample. We can miss important details if we do not apply the correct accelerating voltage. The main reason why field emission gun are so popular is their extraordinary capacity to work at low voltage.

The most important aspect in SEM is the recorded image. For SE-images, the acceleration voltage (Vacc) must be set according with the nature of the sample. In the case of insulator sample, which don't let electrons to pass through sample you must to use low Vacc, but in the case of metallic sample you can chose high Vacc. Also, if you have SEM coupled with EDS, for EDS analysis try high Vacc. 

yes

My take is this, most literature works fail to justify their use of analytical instruments including the SEM and others. They also do not justify choice of operating conditions, which I feel is as important as what is being analysed. I have noticed this in countless number of articles I have read over the years. If you are in doubt of this information, I challenge you to please google any article in which you think not just the SEM was used but any other analytical instrument(s) and, return to me on this platform.

When we are talking about SE not BSE, we have to consider not only the Vacc but also detecting system. With detector which is designed for SE type I even 200kV will give You an image with surface topography. The attached one was taken at 200kV and present carbon fibers on carbon foil.

Interestingly enough, couple years ago at least one of SEM/FIB manufacturers begun embedding optical models of their columns into the operation software, providing estimated beam diameter as function of acceleration/aperture/current/distance conditions to the operator. This IMHO is a great help with at least "beam size" consideration for choosing operating conditions. Contrast mechanisms, charging, depth of signal origination, etc. considerations are still in operator-skill based "art" domain.

Agree on habit affecting lack of active choice, but just using either what the previous person used, or what has been suggested previously, or what has been used in M&M in a paper by somebody else. I'd be surprised if more than 20% of LifeScience operators can articulate why they use any given set of parameters.

On my system (Zeiss EVO 40 XVP), in VP 20 kV is usually the sweet spot for my biological samples for signal yield vs. charging. In HV-SE, 30 kV often leads to translucent thin organic structures, very nicely shown above in the 200 kV image by Thomas Plocinski, which then is no longer a true surface image. Trade-offs are important to recognize.

Beam deceleration has further complicated matters. Don't have experience with it, but is certainly something I will be interested in a new instrument. As Valery Ray said, there is "art" in EM science. There are also the various noise reduction approaches that can mitigate of exacerbate certain effects.

Daniel, did you really mean 20 kV (" the sweet spot for my biological samples for signal yield vs. charging ")? May be it should be 2 kV?

Vladimir: I deal mainly with molluscan shells and also minute orchid flowers. On the instrument I run and having gone through pretty much every parameter combination I can think of, 20 kV, 10 mm WD, 30 Pa chamber pressure, VPSE detector, gives me the best results with coated specimens. Probe current varies between 200 pA (low mag) to about 50 pA (higher mag: cell surfaces). I prefer VP/VPSE for coated specimens as opposed to HVAC/SE because I get more even greyscale with strongly sculptured, 3D specimens. With VPSE and lower EHT I do not get enough signal.

For uncoated samples, I usually go to about 100 Pa chamber pressure. Charge mitigation is rather done by adjusting chamber pressure, then probe current, then noise reduction mode (from line integration to fast frame integration).

There is a wide range of voltages that can be used on an SEM depending on the type of specimen and the resolution needed. This usually ranges from 0.1 to 40 kilovolts (KV), with 10 KV being the most common for biological specimens or heat sensitive samples (e.g. plastics and rubber) and 20 KV for nonbiological

specimens. The higher the voltage the better the resolution but the greater

the heat that is generated on the specimen. Lower voltages may have to be used for delicate samples.

Sorry, but 10 and 20kV are too high voltages. Read discussion again.

I think 10kV is sufficient for this .

Obviously, the question has a deep physical background, but there is a very rough simplistic explanation: the point where the beam contacts with the sample can be approximated as a small volume - lateral size of such "voxel" depends on the beam current and the depth (Z), obviously depends on HV. This voxel is a source of detectable signal and its size limits the practical resolution. Obviously, it should be as small as possible. From this naive geometrical point of view It is self-evident: decreasing of HT (modern FE-SEMs typically operate at 2 kV or even less) is a way to resolution improvement. The de Broglie wavelength limit is still far and this factor doesn't matter in this case. Practical limitations mostly lie in detectors sensitivity.

Just an example of 2 kV FE-SEM picture of biological object (detached part of mouse spermatozoon).

The voltage is an indication of the electrons’ energy content: this will therefore determine what kind of interaction the beam will have with the sample. As a general guideline, a high voltage corresponds with a higher penetration beneath the surface of the sample also known as bigger interaction volume.This means that the electrons will have a larger and deeper propagation within the sample and generate signals in different parts of the affected volume. The chemical composition of the sample also has an impact on the size of the interaction volume: light elements have fewer shells, and the electrons’ energy content is lower. This limits the interactions with the electrons from the electron beam, which can therefore penetrate deeper into the sample, compared to a heavier element. Typical recommended accelaration voltages values for a coating analysis range between 10 and 20kV.

At higher magnifications, much more surface detail is visible in biological samples when using an acceleration voltage between 2 kV and 300 V.