How can Low-e glass have dramatically lower U-Value?

18 posts / 0 new
Last post


I wanted to see if Low-e glass saves energy in the far North (60 deg. N
latitude or more). My thought was to use the same U-value for the
glass, but change the SHGC to account for the difference in solar heat
gain due to the Low-e coatings. To my surprise, manufacturers data for
Low-e glass lists much lower U-values for the same double glazed units
except with a Low-e coating on surface #3.

I'm having a hard time understanding how a coating a few molecules
thick, improves the U-value so much. The Architects in my firm say that
the manufacturers are calculating an improved U-value to account for
energy saved by blocking radiant heat lost (going from inside, out) in
Winter. They surmize this is done because our energy loads and modeling
software cannot calculate radiant heat loses in Winter. I'm not sure
the weather data we use has hourly long wave radiation data that can be
used to determine the available IR heat that can be blocked by the Low-e
coating. I don't think our energy modeling software can account for
radiant heat leaving the building in Winter.

For example,

Pilkington 1" double pane clear glass using air, has a Winter U-value of
0.47 Btu/hr.sq ft F and an SHGC of 0.71

The same Pilkington unit with their Energy Advantage Low-e coating has a
Winter U-value of 0.33 and an SHGC of 0.67

PPG lists similar improvement for their Low-e coating

Is our energy modeling software inadequate to accurately model the
effects of Low-e coating on glass? Both Summer and Winter?

Can we trust that the glass manufactures are giving us improved U-Values
due to Low-e coatings that are valid?

Randy Wilkinson
Spokane, WA

Randall Wilkinson's picture
Joined: 2013-01-07
Reputation: 0

Dear Randy,
What software are you using?

To properly model window coatings you could first use a 2D FEA package such
as Window, then for framing Therm, and for optical you could use Optics.
Then you can use the 2D model results as inputs to 1D software such as
Energy Plus.

Also a nice, quick way to do analysis is to use COMFIN, (in same link) a
graphical UI to E+ that is setup to model windows that have been calculated
with Window/Therm/Optics.

* ??Jeremiah D. Crossett ** | Senior Analyst **| **LEED Green Associate *
*??120 E. Pritchard St. | Asheboro, NC 27203 ?? |
Mobile *

Jeremiah Crossett2's picture
Joined: 2012-12-14
Reputation: 0

By decreasing the emissivity of glass, you improve the glass' ability to
reflect long-wave energy. If the glass doesn't have to absorb the heat,
then it's not going to convect and/or radiate that heat to the indoor
space (which is the method by which it improves the U-value). That's
how a very thin coating can have such a dramatic impact on U-value.

I'm not sure if that answers your question!

GHT Limited
James Hansen, PE, LEED AP, BEMP
Senior Associate
1110 N. Glebe Road, Suite 300
Arlington, VA 22201
703.243.1200 (main)
703.338.5754 (direct/cell)

James Hansen's picture
Joined: 2011-09-30
Reputation: 200

When the building is in heating mode in the far north ( 60 degrees N latitude) the low emissivity coating on the 3d. surface of glass in the glazing unit works in a similar manner to that described below to block long-wave radiant heat transfer from the warm interior of the building to the outside. The long-wave radiant heat is transmitted through the first layer of glass, then reflects off the low-E coating (on the 3d. surface of glass in the glazing unit) back into the heated interior space of the building.

In this way, the radiant heat loss component of heat transfer from the building is blocked and the U-value of the glazing unit with a coating is significantly higher than a unit without a coating. Hope this is helpful.

Judson Brown, PE, P. Eng., LEED AP

Judson Brown's picture
Joined: 2014-07-29
Reputation: 0


Most higher performance windows have a coating that is both low
emissivity and spectrally selective. It's the spectral selectivity -
blocking the parts of the solar spectrum beyond the visible - that
reduces the SHGC. The low emissivity reduces the U-value, as others have


Philip Haves's picture
Joined: 2011-10-02
Reputation: 0

Keep in mind that you want a higher SHGC in the north to maximize passive heat gain.
The framing and glazing spacer selection is important. Argon fill is a must have.

Christopher Jones, P.Eng.
Tel: 416.644.4226 * Toll Free: 1.888.425.7255 x 527

Chris Jones
crollinjones's picture
Joined: 2013-12-12
Reputation: 0

I went into this thinking that the Low-e coating would simply change the
SHGC and not change the U-Values much. I suspected that Low-e coatings
would NOT save energy in the far North since they would block the wanted
solar heat gain. Since the manufacturers are claiming a U-value
improvement from the Low-e coating, It looks like I can't calculate the
savings using our normal energy modeling software. If the U-value
improvement they give is in lieu of a more direct calculation method,
and is accepted by the community (you all), then I will accept and move
on to the next problem.

Thanks to all for your comments.


Randall Wilkinson's picture
Joined: 2013-01-07
Reputation: 0

This is exactly what I mean by asking if our energy modeling software is
inadequate. Maybe a Senior Analyst or Building Scientist can do this, I
don't think I can, or should. If it takes specialty software to model
long wave radiation coming in AND going out, then it seems like the
functionality of Window/Therm/Optics should be built into our energy
modeling software.



Randall Wilkinson's picture
Joined: 2013-01-07
Reputation: 0

The LBL WINDOW program can calculate a U-factor and Solar Heat Gain Coefficient for a particular IGU, i.e., you create a glazing system with the different layers from the Glass Library, make sure the low-e coatings are on the right surface, and the program will do that calculation for the glazing system without a frame. If you want whole window properties, you can put the glazing system into either a generic frame or a frame modeled with THERM, and WINDOW will calculate the U-factor and SHGC for the whole window.

Energy Plus can correctly model a window with low-e glass, and uses algorithms from WINDOW that have been incorporated into the program. You can use a file generated by WINDOW to define the glazing system and frame and insert it into the Energy Plus IDF, or you can built it up natively in whatever interface (or no interface) that you use to run Energy Plus. The DOE2 simulation engine can also correctly model low-e glass.



Robin Mitchell

Building Technology and Energy Systems

Lawrence Berkeley National Laboratory

Robin Mitchell's picture
Joined: 2011-10-01
Reputation: 0


I don't know what energy modeling software you're using, but you certainly don't need a finite-element analysis to do what you've described. Indeed, most of the tools used by regular contributors to this forum take into account all of the following glazing properties:

- U-value (as a function of low-e coating, gas fills, edge spacers, and so forth);
- SHGC as determined by spectral selectivity of coatings, absorption, re-radiation, and convection;
- Visible transmittance as a function of the glass composition

Most of the detailed data for many hundreds of actual glazing products and assemblies of these---needed if you want to use a particular set of inputs to represent this information accurately---can be found in the International Glazing Database from Lawrence Berkeley National Laboratory (LBNL) as viewed within their free Window 6 tool.

Once you have that data, you then need a software tool capable of modeling solar radiation striking your windows (including what passes directly through, what fraction is absorbed, and to which side of the glass this is re-radiated or convectively transferred), heat transfer from the air in the building to the interior of the glass at a given delta-T, and long-wave radiation trying to escape the building through the windows. This capability is common in all the major whole-building simulation packages.

If the software you're using does not account for these fundamental properties of glazing, perhaps you should consider upgrading your software, rather than abandoning your project.

Timothy Moore
Senior Product Manager

T: +1 (415) 983-0603
T: +44 (0) 141 945 8500
timothy.moore at

Integrated Environmental Solutions Limited. Scotland registration SC151456
Helix Building, West Of Scotland Science Park, Glasgow G20 0SP

Email Disclaimer

Timothy Moore2's picture
Joined: 2011-10-02
Reputation: 200


I think you're being misled in a way. ALL building simulation programs model long-wave
radiation between the building surfaces and the environment, because otherwise you would
get erroneous results, a case in point being night-sky radiation that causes roofs to be
significantly colder than the outdoor air at sunrise. How different programs handle
long-wave radiation varies, but that's more an issue of modeling methodology, whether to
combine the radiative with the convective or calculating them separately, what temperature
to assume for the
environment (ground, sky, air, etc.), etc.

As for the LBNL Suite of window simulation software (Window/Therm/Optics), I don't know of
anyone except the NFRC Simulation Laboratories that use all three, and only for the
purpose of getting an NFRC rating of a specific product. Outside of that context, the
most I've seen people
do in building energy simulations is to obtain or create a "Window-4" file using Window
(but not Therm or Optics) and then import that into their building energy software. Even
there, the main advantage is to get better representation of the angular-dependent
properties of the window.
As far as capturing the long-wave radiation, inputting the U-value from an NFRC Rating or
a Window-4 file should work fine.


Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 108D
Moraga CA 94556
yjhuang at for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Joined: 2011-09-30
Reputation: 406

Thanks Joe,

So you are saying that use of an improved U-value is a valid work-around
in energy modeling software that can't or doesn't model long wave
radiation in and out. What if I put that improved U-value in there and
the program then calculates the effect of long wave gains and
losses...would we then have double accounted for the radiation effects?

Since I recognize you as also an expert at simulation weather data, do
our typical hourly simulation weather data files contain sufficient
information to model energy losses and gains from long-wave radiation?
Especially gains from the Sun?


Randall Wilkinson's picture
Joined: 2013-01-07
Reputation: 0


When architects say things like, "our energy loads and modeling software
cannot calculate radiant heat losses in Winter", that is code for "why is
reducing WWR showing better results than any other enclosure strategy".

On Monday, July 28, 2014, Randy Wilkinson


Arpan Bakshi

646.704.2880 mobile
212.298.9352 office

NYCCTgreen's picture
Joined: 2011-09-30
Reputation: 0


I'd like to dissuade you of the idea that a U-factor that includes long-wave radiation is
some sort of work-around. In fact, it's the standard definition of U-factor as applied to
windows because of their high conductivity. Actually, the more I think about it, it would
be the conduction-only U-factor that would be misleading. I suppose it is possible that a
program might use such a U-factor, then also do a external radiative exchange, and thus be
double-counting, but that seems unlikely because to do that you would have to solve for
the external surface temperature, which would require a detailed heat balance, etc.

Sure, standard weather files have all the information needed to account for long-wave
radiation, because during the day, the long-wave radiation from the sun is included in the
reported solar radiation, which is all spectrum, while the long-wave radiation exchange
with the ground, air, and sky are all calculated using assumed temperatures for each and
generally minimal. During the night, the main long-wave radiation exchange is with the
night sky, for which there are various algorithms to estimate the night sky temperature,
depending on the atmospherics, clouds, etc.


Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 108D
Moraga CA 94556
yjhuang at for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Joined: 2011-09-30
Reputation: 406

There is my problem. For decades I've been trained to understand that
energy exchange thru glass contains 2 components: 1) Conduction -
Calculated using a U-Value with units of Btu per hour per SF per degree
F., and 2) solar - calculated using the sun's energy output, adjusted
for latitude, shading factor for glass, tinting, drapes, etc.

The U-Value is only for definition, and is confirmed by
looking at it's units (per degree F). Now a few molecules of metallic
coatings added to one surface adds to the U-value? That's the part I
find hard to believe. Or at least I did until today.

I can accept that an improved U-value can be determined that would make
it easy to calculate the overall equivalent heat transfer, averaged over
a season, including the effect of long wave infrared radiation from the
Low-e coatings. That would make less sophisticated simulation programs
be able to get better results. I guess I just need to verify that the
improved U-value given is an NFRC value that has been determined using
LBNL's Window software.

Thanks to all for helping me understand. As others pointed out, E+ can
accept a Window glass definition file and can also allow you to define
glass parameters using layers. I better figure that out next.


Randall Wilkinson's picture
Joined: 2013-01-07
Reputation: 0

I have been following this discussion and I?m still a little unclear. As I understand it, the U-value for windows from manufacturers includes the long-wave radiation as well ? accounting for the effect of the emissivity of the window on the heat transfer and linearizing the radiation component at the winter and summer temperatures to create a combined radiation/conduction U-value.

I am not convinced that using one of these effective U-values in certain softwares will not double count the effect of long wave radiation. For example, when I look at eQUEST?s inputs for simple window definitions, it has fields for:

- Shading Coefficient: to get the short wave solar radiation

- Glass Conductance (U value)

- Visible Transmittance: for daylighting calculations only

- Outside Emissivity: from the DOE2 help: Outside surface IR emissivity of glazing. The default value (0.84) can be used for uncoated glass. Glass with an outside surface metallic coating would have OUTSIDE-EMISS between about 0.03 and 0.4 depending on the type of coating. Used to determine IR radiation leaving the glass and absorbed by the glass.?
So it seems to me that, at least in eQUEST, the IR heat transfer into and out of the room is calculated separately based on the user-input emissivity of the glass. If the U-Value included the effect of the emissivity, wouldn?t that be double counting the effect of short wave radiation? Someone who knows more about window ratings or modeling software could please help clarify.


Justin DeBlois | Energy Analyst
Justin.DeBlois at
Voice 646-253-9000

Anonymous's picture


Some of the commonly used simulation engines do account for the effects of low-e glass on window surfaces (for example, Energy Plus and ESP-r). That is to say that these engines estimate the long wave radiation from interior surfaces (and devices) to the interior surface of the glass and radiated through the various glass panes and interstitial gases to the sky/surroundings.

Low-e coatings effect the overall U-value of windows because of their effect on reflected long wave radiation. Which low-e coating is most appropriate for the North is important since the coatings will affect both the SHGC and the U-value, depending on their type, thickness and the surface they're applied to.

The results of a recent study of a Northern Canadian city (in NWT) concluded that hard coat windows, despite their higher cost and relatively higher U-value make sense in terms of both energy savings and cost-effectiveness.

Jeff Blake, P.Eng.
Sustainable Buildings & Communities | B?timents et des collectivit?s durables
CanmetENERGY | Canmet?NERGIE

Blake, Jeff2's picture
Joined: 2011-10-02
Reputation: 0

My own idea in equest was to use the four detailed user inputs that jd508 talked about 2 posts ago for the glazing.

The question is this... If you model clear glass in equest but tweak the emissivity of the system to allow for the heating benefit of the low-e coating in winter, you are effectively modeling the low-e coating, correct?

The U-value would still need to include the low-e coating if it applies but typically the U-value from the manufacturer would account for it, correct me if I'm wrong.

A clear glass window typically has emissivity of .83 (which is the default in equest I believe). The low-e coating makes the overall system more efficient in retaining heat within the structure in far north winter conditions. Even if it is only minimal it still plays a role in winter. the emissivity of a low-e glazing depends on the manufacturer too.

I recently read that some windows have emissivity of .04 or something insane provided by having a special kind of low-e coating.

However, If you are able to figure out the glazing emissivity created by having the low-e coating ( for example, .4) and you type that in under the glazing properties in equest, you are effectively manually modeling the effectiveness of the low-e coating in equest without needing to do anything else.

The default emissivity of .83 in equest just is not accurate for low-e coated windows. That's for clear glass only.

Personally, I was able to call my supplier and get the SHGC, VT, U-value etc.. for the glazing and I used all of that data in equest for the proposed model (for a project we are submitting for LEED cert).

Regarding U-values (glazing vs overall) , you should check with the manufacturer. They typically have tables that can help. If you only have the U-value at center of glass (COG) you can use their bell curve to figure out the overall U-value of the system - and vice versa. Also, the overall window dimensions may play a part in determining the overall system U-value.

That's pretty handy for when it comes time to do those oh-so nasty eap2 tables because you need the overall U-values of the baseline and proposed window systems for the table inputs.

I still don't think equest can give you a window-system overall U-value based on all the user inputs.

In my opinion, The big benefit in far north isn't the U-value provided by the low-e coating but the emissivity factor, specifically. Because in cold weather climates you want that heat to be retained a bit more efficiently by having the low-e coatings.

I don't see how a low-e coating is going to affect U-value much at -40F or -50F but if the manufacturer-tested result is X, I'd go with that. At these extreme cold temps, thermal breaks, frame type/size would play a much bigger factor on the overall system (which in equest are modeled separately from glazing in the window properties).

I should also mention that if you are modeling per LEED tables 5.5-1 thru 5.5-8 your baseline factors are right there in black and white. if the U-value in those tables is lower than the U-value of the proposed model, you use the U-value in the table if I remember correctly.

-Chris in Alaska

Chris Baker2's picture
Joined: 2014-03-25
Reputation: 0